Suspension (and new team) Guidance

[Nbruno]

Hey guys, I know most of teams are busy finishing and testing cars, but I could use some help with our suspension design for next year. Some background first...

I'm on the Syracuse University FSAE team. We just got the team back together after a 10 year haitus and we have started design for next year's car. We started with choosing a wheelbase (73in) track (54in front 50in rear)and our tire (20.5 x 7.0-13 Hoosier). We also obtained OptimumK to assist in suspension design and analysis. We just put the suspension of the 10 year old car into the program and have been playing around with it a bit in between studying for finals. I must say I am very impressed with the ease of use and analysis tools of the software. However, it has exposed a gaping hole in my design plans...I realized the program is too powerful for my current knowledge, and I don't feel that I can fully take advantage of it.

I have been reading a lot since I joined the team this year. SAE papers, Alan Staniforth's racecar suspension book, Carol Smith. I worked with one of the best chassis setup guys in the US setting up F2000 cars last summer and I've had some contact with ANZE suspension. I have a good grasp of camber, caster, toe, KPI, MR etc...but the problem is I am not sure how all of these different components come together to form a good suspension design. I've gotten some tidbits (ANZE told me no camber change through bump) but I feel like I would still be stabbing in the dark with whatever design I choose. As I'm sure many of you can relate to, I need a reason behind my design.

So I guess what I'm looking for here is a starting point. What's a good degree of roll to test properties of the suspension at? What am I looking for in terms of camber change curves? Bumpsteer? Basically, in terms of suspension parameters, what makes a "good" suspension design and how can it be tested for? Any help would be greatly appreciated, even if it's more suggested reading! http://fsae.com/groupee_common/emoticons/icon_smile.gif

[Nbruno]

Hey guys, I know most of teams are busy finishing and testing cars, but I could use some help with our suspension design for next year. Some background first...

I'm on the Syracuse University FSAE team. We just got the team back together after a 10 year haitus and we have started design for next year's car. We started with choosing a wheelbase (73in) track (54in front 50in rear)and our tire (20.5 x 7.0-13 Hoosier). We also obtained OptimumK to assist in suspension design and analysis. We just put the suspension of the 10 year old car into the program and have been playing around with it a bit in between studying for finals. I must say I am very impressed with the ease of use and analysis tools of the software. However, it has exposed a gaping hole in my design plans...I realized the program is too powerful for my current knowledge, and I don't feel that I can fully take advantage of it.

I have been reading a lot since I joined the team this year. SAE papers, Alan Staniforth's racecar suspension book, Carol Smith. I worked with one of the best chassis setup guys in the US setting up F2000 cars last summer and I've had some contact with ANZE suspension. I have a good grasp of camber, caster, toe, KPI, MR etc...but the problem is I am not sure how all of these different components come together to form a good suspension design. I've gotten some tidbits (ANZE told me no camber change through bump) but I feel like I would still be stabbing in the dark with whatever design I choose. As I'm sure many of you can relate to, I need a reason behind my design.

So I guess what I'm looking for here is a starting point. What's a good degree of roll to test properties of the suspension at? What am I looking for in terms of camber change curves? Bumpsteer? Basically, in terms of suspension parameters, what makes a "good" suspension design and how can it be tested for? Any help would be greatly appreciated, even if it's more suggested reading! http://fsae.com/groupee_common/emoticons/icon_smile.gif

[js10coastr]

"Race Car Vehicle Dyanmics" is a good book to start with... Your degree of roll will depend on your springs along with your geometry and performance... Essentially you need to take a guess, a shot in the dark, or use tire data to make an educated guess on what the car will actually do performance wise, design the suspension around that and then re-iterate (also re-iterate after building and testing).

[RyMan]

Disclaimer: This will likely be a discontinuous series of ideas.

A good suspension keeps the tire happy and in contact with the ground as frequently as possible. Tires do not enjoy positive camber. Drivers do not enjoy bumpsteer so keep it as close to zero as you can. Drivers also do not like too much steering effort or too little (caster, pneumatic/mechanical trail play a role). KPI and caster impact camber change through steering. Use as much of the travel in your damper as you can. MR drives how much your wheel moves relative to your spring/damper (1:1 is what we shoot for). Make sure you can keep your car from hitting the ground in bump. We see what happens in roll up to two degrees or so, but that depends on your roll rate which depends roll resistance if i'm not mistaken. It's been a little while since i've read through that stuff.

Grab a copy of Race Car Vehicle Dynamics if you don't already have it. I'll let the community at large fill in the gaps I left. Time for me to get back in the damn shop.

[bminus]

Another book I'd recommend (especially if RCVD seems a bit intense at first) is "How to Make Your Car Handle" by Fred Puhn. There are a few basic principles that I think are spelled out pretty well in that book.

There is no one way to do this. The one constant you will find in every book is that they all mention compromise. There are gives and takes with every aspect of suspension design. The best designer will know exactly what type of car he/she is attempting to create for the driver, how and why it will get there, and how to adjust if it doesn't go perfectly the first time (and it won't... I promise).

[Barky]

Carroll Smith's books are much easier to read an comprehend than RCVD, and just as good for 99% of what you need to know at this point, imo. Tune to Win is what you want (and it sounds like you've found it) if you're working on your suspension, although Engineer to Win should be required reading for everyone working on the car.

"Any suspension will work if you don't let it..."

[sbrenaman]

I prefer Think Fast. It has a section in there about handling parameters/interactions.

[Kevin Hayward]

Nbruno,

The best thing to keep in mind is that you will definitely not get it spot on in design. For your early cars try and make sure you build adjustability in things like ackermann, caster, enough camber and preferably KPI and mechanical trail. That way if it ends up a real mess then you have room to move.

That being said here are a couple of points:
- These cars can be very sensitive to steering geometry (caster, trail, scrub, and ackermann)
- Bump steer is very easy to stop in design so no reason not to do it
- Different tyres (e.g. Hoosier vs Goodyear) like very different amounts of camber.
- Many cars run conventional FVSAL and roll centre heights with a lot of success. (1500-3000mm SAL Front, 1000-1500mm SAL at rear, 0-50mm RC front with rear about 25mm higher than the rear)
- Sensitivity to SAL and RC is not super high and these should be considered ranges you can work in to get good packaging.
- Normally a car will have 0-5% more roll moment distribution than weight distribution at the front. However due to large steering angles you are most likely going to need at the low end of that or even less (than 0) to avoid the understeer.

Hope some of this helps. There is a lot posted about suspension on these forums that is a great source of reading.

Kev

[murpia]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Kevin Hayward:
- Normally a car will have 0-5% more roll moment distribution than weight distribution at the front. However due to large steering angles you are most likely going to need at the low end of that or even less (than 0) to avoid the understeer. </div></BLOCKQUOTE>
Hi Kevin,

How are you defining these distributions?

Normally if a car is weight biased e.g. 40% front axle and 60% rear axle, I would expect to see a roll moment distribution around 60% front and 40% rear (i.e. the relationship is 'flipped' front axle / rear axle). +/- 5% of that range is a good tuning range.

I'm sure this is what you meant by your post, but taken as read it would imply a 45% front roll moment bias for that car which seems very low to me.

Regards, Ian

[js10coastr]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by murpia:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Kevin Hayward:
- Normally a car will have 0-5% more roll moment distribution than weight distribution at the front. However due to large steering angles you are most likely going to need at the low end of that or even less (than 0) to avoid the understeer. </div></BLOCKQUOTE>
Hi Kevin,

How are you defining these distributions?

Normally if a car is weight biased e.g. 40% front axle and 60% rear axle, I would expect to see a roll moment distribution around 60% front and 40% rear (i.e. the relationship is 'flipped' front axle / rear axle). +/- 5% of that range is a good tuning range.

I'm sure this is what you meant by your post, but taken as read it would imply a 45% front roll moment bias for that car which seems very low to me.

Regards, Ian </div></BLOCKQUOTE>

The "5%" has been a "Claude - ism"... I stuck with that number until I worked with some FWD cars that had 60% front WD. To get the car to handle we needed something close to 60% rear roll stiffness (same tires front and rear).

[wil oberlies]

Nbruno, you have a pm...

[wil oberlies]

I think one of the most significant things to be cognizant of is the compromises that are necessary when designing a suspension and setting up a car for a given event. The choice of springs, for example, will set the average ride height of the car, its attitude, and the spring split will affect the lateral balance. Higher rate springs will limit suspension travel (hiding 'bad' geometry to a degree), and increase the rate of heat input to the tires. They (high rate springs) also cause more variation in contact patch load, reducing mechanical grip relative to a more softly sprung car. Softer springs will provide more grip over surface irregularities, but may result in a car that feels unresponsive to the driver.
Shocks and springs both determine how energy is stored and dissipated in the suspension, and the type of shock and its characteristics will affect load variation at the contact patch, and driver feedback things such as steering time constant.
All of the above is basic stuff, the purpose being to illustrate the types of compromises one is faced with. If you know something about the course you'll be running on, and can get some tire data, I think you can go about making some intelligent decisions...

[Kevin Hayward]

Ian & Justin,

I did borrow a "Claudism", and in my haste I poorly defined it. Ian's comments are correct (as I understand Claude's simplification).

When working on vehicles calculating the wheel loads due to cornering (and combined handling states) is much more useful, and not a big step from having the roll moment distribution defined. I should have suggested you do this as part of the design process.

On the cars with higher weight distribution differences you are also accounting for different tyre sizes which complicates things a little. I would only recommend any of these rules of thumb as a starting point. At some stage in the design you will need to look at everything together and apply a common sense filter. If you had applied the poorly worded rule to a heavily rear biased car during corner you would have seen a massively over-worked rear tyre. If it was the same size as the front tyre bad things (or fun things) would happen.

Kev

[Nbruno]

First of all, thanks to everyone who replied.

I think the statement that helps the most is " keep the contact patch as large as possible through travel" something so obvious that got washed away in the noise of all the technical talk for me. Then all the parameters are simply tools to achieve that end goal.

[Goost]

Just from my limited experience, but you should know sometimes the contact patch size doesn't matter so much if you can't get the whole thing hot. May not be a problem with the Hoosiers and other bias ply tires that most teams seem to use. Definitely matters with our radial Michelins though.
Another important thing is keeping all four wheels in contact with the ground; not quite as easy as it sounds sometimes.
You seem to know all the general parameters, best thing I could suggest is joining the TTC and having access to that tire data. In the end it's all about the tires.

[exFSAE]

Think top-down, not bottom-up. It's very easy to get lost in a myriad of suspension parameters and kinematics tools without really knowing what you're trying to achieve, in which case it becomes guess-and-check rather than design. A good engineering DESIGN process meets the goals you specify, which may or may not involve using fancy analysis tools. As Geoff from RMIT alluded to in his big thread, starting simple is the way to go. A point vehicle model, or bicycle model, will probably tell you more valuable stuff sooner, than going into OptimumK or ADAMS or whatever if you're just starting out.

Some things to consider...

- What's most important on these cars? Grip? Response? One more than the other, or the same? Should hint at what you want out of your tires.

- For grip, what's more important - lateral or longitudinal? Should hint at what you want in terms of kinematic rates and camber control, be it very good in roll or very good in heave and pitch.

- Longitudinal grip, what's more important driving or braking? Points to what kind of static nose weight you might want.

Gotta start thinking about this stuff at a high level. If you don't know the answers to those questions (e.g. what do YOU want the car to do) it's pointless spending time and money on analysis tools.

Once you CAN set those goals, the specifics are easy. Know if you want better camber control in bump vs roll? Pretty much sets VSAL. Figure out what kind of jacking and/or roll center heights you want? That and VSAL length pretty well defines where your A-arms are going to be, given some geometric constraint of the wheels and chassis. It really starts falling into place, you don't even have to think about it - takes care of itself.

[js10coastr]

IIRC, people in the US (IE US automakers) didn’t start paying attention to handling/suspension design until the 1950’s or 1960’s… so really it’s a pretty “new” science that has only had two or so generations of engineers going through it. There is no one “right suspension”, but everything in it is all a bunch of compromises. It takes a (very long) while but you’ll start to understand it all and put it together. Once you understand the tradeoffs (ie the camber change in a short v long FVSAL), the design process starts to get really fun. I’ve seen a handful of “new” teams get caught up in the theory (and scared of getting it wrong), and get stuck on “roll center lateral movement under corner entry” or something like that… keep in mind you still have to build the damn thing! For a first year, I’d just get everything “within reason”, and then test, tune, and learn from it (you designed in that adjustability that was suggested by Kevin right?). If you have access to that F2000, maybe take some measurements… run it through a suspension program and see what it does(?).

Kev! Great to see you on here! I wish I could in someway tangibly share the joy’s that I’ve had over the last year at the track! I feel that all the success I’ve had is due to the people that I worked with at “the G”. There is no way I’d be where I am now had you not been there. It’s a crazy story on how I got on with Black Swan too! Hope you, Jodi, and the kids are doing well!

[woodsy96]

Some good advice. I particularly like the "Top down" design philosophy, but more importantly the getting it to work first and making sure you don't have interference issues (like I did. My bad) http://fsae.com/groupee_common/emoticons/icon_smile.gif.

You will learn a lot more from having a running car and changing things than sitting at a desk for 6 months trying to figure it out. Allowing equal time for design and testing throughout the year is a fantastic starting point.

There are a couple of things that have been missed that I feel are quite important.

1) Stiffness, in particular camber (front and rear) and toe (especially rear). If your car is too sloppy than making a change on the car won't make a change on the track. You won't learn anything about the design decisions you did make (whether they were right or not - this will be important for future cars your team builds to learn from) and you won't reinforce theory with practice. You also won't be able to tell if your massive adjustment range you are going to build into your car (hint hint) will make any difference.

2) Allow for lots of steering, maybe 30+ degrees on the inside wheel. This allows for lots of slip angle on the front and lots of room to get out of trouble if your driver makes a mistake.

It seems like you are on the right track though. Good luck.

[murpia]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Kevin Hayward:
Ian & Justin,

I did borrow a "Claudism", and in my haste I poorly defined it. Ian's comments are correct (as I understand Claude's simplification).

When working on vehicles calculating the wheel loads due to cornering (and combined handling states) is much more useful, and not a big step from having the roll moment distribution defined. I should have suggested you do this as part of the design process.

On the cars with higher weight distribution differences you are also accounting for different tyre sizes which complicates things a little. I would only recommend any of these rules of thumb as a starting point. At some stage in the design you will need to look at everything together and apply a common sense filter. If you had applied the poorly worded rule to a heavily rear biased car during corner you would have seen a massively over-worked rear tyre. If it was the same size as the front tyre bad things (or fun things) would happen.

Kev </div></BLOCKQUOTE>
Maybe we should spin this out into a separate thread?

Anyway, here's what I (think I) know, for a RWD racecar, hopefully explained clearly:

Weight distribution is the number 1 factor for the other aspects of your car (tyre contact patch distribution, roll moment distribution & aero distribution).

A simplistic view would be that a 50/50 weight distribution would use equal size tyre contact patches, equal roll moment distribution and equal aero distribution.

Rarely do you get the option of 50/50 weight distribution so for arguments sake lets analyse 40/60 front/rear which is more typical of a mid engined racecar. Where might you start your design & track tuning process?

A first option might be to go with 40/60 tyre contact patch distributions, especially since with RWD you can benefit from the additional traction.

Then, if you have aero, biasing that 40/60 front/rear would make sense given the need to keep the axle vertical loads in line with the weight and tyre contact patches, through the vehicle speed range.

That leaves roll moment distribution to decide on. Up to now we have been balancing the required axle lateral forces, for steady state cornering, against the weight distribution. I.e. we can expect 40/60 tyre grip and 40/60 aero balance to give us axle lateral forces split 40/60 also. Taking moments around the CoG we see that 40/60 lateral forces balance a 40/60 front/rear weight bias.

What do we require of the roll moment distribution? It's effect is to bias grip to one axle or the other in proportion to lateral acceleration. It only does this because tyres produce slightly less lateral grip as vertical load increases. E.g. double the vertical load and you might get 1.8x the lateral grip. Naturally there are complications, and whole books are written about them, but that's the essence of the situation.

So, for our 40/60 racecar we might try 50/50 roll moment distribution, since we think we have already compensated for the 40/60 weight bias with our 40/60 tyre and aero biases. That should result in a 'neutral' racecar, which often is not what we want: Most drivers prefer / need some inherent understeer in the car in order to receive adequate feedback from the steering. Also, any analysis of stability would show that a truely neutral car cannot easily be driven, even in a straight line, without active control from the driver. Those factors therefore require us to build in some understeer in proportion to lateral acceleration, achieved by a roll moment distribution which biases grip toward the rear axle.

In the mathematical conventions normally used, this means a shift from 50/50 front/rear roll moment distribution towards a numerically greater front / lesser rear value. Whether you go all the way to 60/40 front/rear or not depends on other factors not discussed here, such as cambers. But I would suggest it is a good place to start.

Note that in all cases I refer to a definition of roll moment distribution which INCLUDES any geometric effects caused by instant centres at any other height than the ground plane. Call this 'roll centre height' if you prefer, but the effect needs to be accounted for in the maths. Your calculation should also include the tyre vertical stiffnesses. If you have pushrod loadcells, and ground plane instant centres, it is actually pretty easy to measure. Likewise if you have corner weight scales, a wall and a ratchet strap, it can be measured. (See Billy Shopes pages on w w w . racetec . cc / shope).

Regards, Ian

[exFSAE]

...contact patch distribution?

[murpia]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by exFSAE:
...contact patch distribution? </div></BLOCKQUOTE>
Maybe not a very good description, but I lack a reference for anything better.

It was an attempt to summarise the widely observed effect that wider tyres have more grip. However I doubt there is an exact width -&gt; grip relationship, but I would expect it to be possible to measure the effect on car balance of changing relative tyre grip front / rear by fitting different width tyres.

Maybe someone with access to the TTC data can wade in and suggest a suitable relationship? Or maybe that's a track test someone's done?

If your car had a 40/60 front/rear weight bias, would you think it worth trying a 40/60 front/rear tyre width bias? I would.

Regards, Ian

[exFSAE]

Ahh I see. Gets a bit more involved. With the same tires all around I'd probably lean toward 55/45 or generally low nose weight for a RWD racecar. Maybe even as far as 60/40. Going with bigger tires in the rear.. that would change.

Tire dimensions are a bitch. Width and OD do interesting things.

But we digress. In any advice, my advice is still to only get as "advanced" as you truly understand in terms of dynamics. There's a lot to be said for building a car that's RIGID and robust with some adjustability, and getting in your test and tune days. You'll kick the ass of a lot of teams stuck on intellectual masturbation.

[Rotary Sprocket]

This has already been said but i'll say it again. Don't get wrapped up so much in the theory of your suspension. No matter how hard you try your calculations are never going to match what the car is actually doing. There are too many variables that are constantly changing. The most important thing is to get your suspension close, and then build it and test it. It's so much easier to test theories when you can make a change to the car (as said before set the car up for a wide range of adjustments) and see what it does on the track. There's a lot of good advice on here and it sounds like you're off to a really good start. Just remember theory has it's limits and it's only meant to get you in the ballpark. The most important thing is to build the car and test, test and then test some more.

[Pete Marsh]

"So, for our 40/60 racecar we might try 50/50 roll moment distribution, since we think we have already compensated for the 40/60 weight bias with our 40/60 tyre and aero biases. That should result in a 'neutral' racecar,"


So, to join the thread highjackers.

I call BS here.
If the axles have relatively the same tyre loads on them, due to their tyre size difference, then I believe we would need to distribute Roll Moment in proportion to their static distribution, so as to impart the same PERCENTAGE load variation on each pair, thereby maintaining the balance as it was. I agree some under steer is in order.

This is in agreement with Claude's notes as I understand them. Perhaps if the author is lurking somewhere....

Pete

Oh, and to the OP. Don't forget to keep your motion ratios as constant as possible, and the inevitable errors in the same direction and same magnitude front and rear. Also make very sure the as built car accurately reflects your design. This is often overlooked, and incorrectly adjusted rockers can easily ruin the cars behaviour.

[flavorPacket]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Pete Marsh:
I call BS here.
If the axles have relatively the same tyre loads on them, due to their tyre size difference, then I believe we would need to distribute Roll Moment in proportion to their static distribution, so as to impart the same PERCENTAGE load variation on each pair, thereby maintaining the balance as it was. I agree some under steer is in order.
</div></BLOCKQUOTE>

Roll moment distribution is not load transfer distribution...

[Pete Marsh]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by murpia:

Note that in all cases I refer to a definition of roll moment distribution which INCLUDES any geometric effects caused by instant centres at any other height than the ground plane. Call this 'roll centre height' if you prefer, but the effect needs to be accounted for in the maths. Your calculation should also include the tyre vertical stiffnesses. If you have pushrod loadcells, and ground plane instant centres, it is actually pretty easy to measure. Likewise if you have corner weight scales, a wall and a ratchet strap, it can be measured. (See Billy Shopes pages on w w w . racetec . cc / shope).

Regards, Ian </div></BLOCKQUOTE>


I think this definition is. Lateral load transfer distribution if you prefer.

I was questioning the "flipping" of weight distribution to get Claude's magic number. I believe you add 5% to the front, for your initial understeering start point, and that it works well.

As for the front drivers, this theory only applies with all 4 tyres on the ground, and as power down rules, we get to see plenty of 3 wheeling.

Pete

[exFSAE]

You can get fancy and include geometric effects and tire rate... but bear in mind tire rate is going to change with:

(a) Speed
(b) Inflation pressure
(c) Vertical load
(d) Camber
(e) Lateral force
(f) Longitudinal force
(g) Usage history (different on lap 5 than lap 4)
(h) Temperature
etc etc...

Good example, can make your model increasingly complex, but not necessarily more accurate given the things that you can't always nail down.