Design Problem

[LuisN]

I'm designing one FSAE car for the first year in my University, and I'm responsible for the vehicle dynamic. But as our first year we have lack of information about doing this.

I have one doubt, the camber gain during the steer caused by caster angle is one thing bad?

And, if is bad how I can compensate this?

And more, how much toe gain caused by roll is bad, and how much is acceptable?

Sorry for my English.

[Claude Rouelle]

LuisN,

It would be nice if you would introduce your self. Just a matter of courtesy.

Your English is fine: everybody will understand you.

I will answer your question by another question: what is the ideal camber (toe) and how do you determine what the ideal camber(toe) is?

[Jonny Rochester]

Good questions.

Caster is designed into the front of the car on purpose, and one of it's main reasons is to change the camber angle when turning (dynamic camber change). It's a good thing. If the wheels have 0 camber when straight, when turning the outside wheel will get neg. camber, and the inside wheel pos. camber (if you also have caster). With some roll, and some tyre flex, these angles should keep a good footprint of the tyre on the ground giving good grip.

Try about 5 degrees caster and go from there. Some static camber may also work for you. Also remember KPI changes these things also.

It's ideal to design a suspension system where you can adjust these angles, with either shims of a thread.

Normal cars do have front toe change with roll, designed on purpose to give some understeer. With race cars we try to design or adjust so we have no bump steer, or toe-change with roll. If you do accidentally have some toe-change with roll it can be awful, or it could be acceptable. If with roll, the toe change increases the steering angle on both wheels, this will be awful and make the car twitchy and had to drive (dangerous even). If with roll, the toe change lessens the steering angle (real cars have this), then this will be predictable and controllable however you will get slight understeer. (As a band-aid for this you can have static toe-out).

[LuisN]

I'm Luis Nigro, student of mechanical engineering in Brazil, UNESP. It's our first year trying to compete in FSAE, and I'm responsible for the dynamic of the vehicle.

Claude Rouelle,

I understand by ideal camber/toe as the state of the tires that generates maximum grip for determinate path. Like acceleration, endurance, skid-pad, and slalom for FSAE cars.

How to determine those values is matter of tires data, and forces on suspension geometry to make the car handle as much lateral forces as possible.

Due these, I realize that caster angle generate some camber gain on corners and I don't know at with point it is valid and good for the car. Furthermore the fifth link on both rear and front suspension are responsible for toe gain in roll, these toe gain that I don't know how to make it a acceptable for the driver, and in FSAE competitions.

More specific, the values that make the car unstable. That only in practical you know.

Jonny Rochester,

Thanks for the advice, I'll try to simulate the dynamic comportment of my steering geometry as well my rear and front suspension. But perfection in number building the car and so making the car in the right spot for all corner aren't possible, my question is how much of error in FSAE cars are acceptable. And some band-aid for those errors.

Thanks for all

[LuisN]

I don't know if I made my doubt clear...

If anyone can comment on this... many views but low answers.

Thanks

[Jonny Rochester]

The front wheel alignment settings have a fairly big window in which the car will still perform as a car. The settings are your choice, so small deviation from that is not an error, but just your choice.

You can have the caster anything from 0 to 15 degrees. Try 5 degrees to start with.
Start with camber at 0, then experiment with up to -3 degrees. You may end up with -0.5 static camber for example.
King Pin Inclination could be 0 or 15 degrees. You need to balance this with scrub radius.
Scrub radius should be positive, maybe 10mm. 40mm maybe a bit too much.

Also note that when the car rolls, your suspension may cause the camber to change also. In total it's too many angles for a person to imagine 3 dimentionally and dynamically, which is why people use software. But if you don't have time stick to some general ballpark figures.

With toe change, it is not too hard to have suspension geometry so there is theoretically zero toe change with bump. If the 5th link is on the same plane as 2 other links. Then the real toe change will only be due to flex and movement in joints.

[Jonny Rochester]

Just a word of warning, the older people on here hate to give answers to new people. So don't get up your hopes.

However I'm a bit different. I'll have a go.

[BillCobb]

Given the data for the tire you (or someone else) have chosen, the 'best' caster setting would probably be the value which shifts the peak tie rod load gradient (net tierod force per g lateral acceleration) up to the peak net (Left + right) side force value. Because tire's Mz peaks out before Fy, does, a driver reading Mz induced steering force may fail to sense and hence operate the race car at it's maximum lateral acceleration. To reduce this enigma, caster is added to shift the peak Mz slip angle gradient upward. This leaves the tierod load gradient positive at peak lateral g. The negative side of this will be increased steering torque ('effort') and probably an increase in total vehicle understeer because of the higher loads on the steering system.

Also, caster (i.e. the derivative of camber by steer) effect is small at high speed because the road wheel steer angles are small, the effect on cornering is also small unless the tires have fairly high camber stiffness (bias ply tires do so). If the tires are true radials, it's a small effect. At very low speed, where high road wheel steer angles are large, the camber gain could be a factor, again depending on the tire camber stiffness and the vehicle's level of nonlinear range understeer. You may want to calculate and measure the suspension deflection camber (camber change per unit of lateral force) and the chassis' camber by roll fraction, because these two suspension parameters can degrade the car faster than any caster effect, IMHO.

While you are at it, look up Angelo Nutti (from Brazil) on LinkedIin, contact him for further advice and send him my regards.

[BillCobb]

And us old people prefer to discuss the journey rather than the destination...

[Pat Clarke]

Luis,
Look at this blog article. It may help you.
https://www.formulastudent.de/academy/pats-corner/advice-details/article/overview-for-new-teams/

Pat Clarke

[LuisN]

Thanks for all.

I'm in the middle of the journey, but as our first I made some starting mistakes. I'm also asking for advice to avoid some other mistakes that are possible.

================================================== ============================================

My problem was that I add too much caster angle that is adding too much camber in turns (15 degrees). As I never made one car I don't know this is a bad thing.

In addiction my car have 0.5 degree of toe variation dynamically on the roll, and again I don't know if it is a bad thing.

================================================== ================================================== ===========

BillCobb I wasn't able to find Angelo Nutti on LinkedIn.

Thanks.

[BillCobb]

I know Angelo very well.
https://www.linkedin.com/profile/view?trk=contacts-contacts-list-contact_name-0&id=25154044

Daniel, less so:
https://www.linkedin.com/profile/view?trk=contacts-contacts-list-contact_name-0&id=1101817

But, Ask Daniel for a visit and tour. Then you can discuss ALL of your concerns and issues. Perhaps job opportunity possibilities as a Co-Op Student.

Rather than trying to find University courses in Vehicle Dynamics, nothing like a Boot Camp Co-op session to focus your current and future goals and aspirations.

Good luck. There is something not right about your caster and roll-steer results.

[LuisN]

Thanks BillCobb I'll try to communicate with they

[Z]

I'm in the middle of the journey...

Luis,

Your journey into Vehicle Dynamics is only just BEGINNING. :)

There is a mountain of useful information on this Forum, and also on the wider Interweb. But, most unfortunately, the useful stuff is hidden under a thousand giant mountain ranges of distractions, from trivial chit-chat, through poorly-phrased, and so misleading, advice, all the way to outright bulldust.

It is your job to somehow distinguish the good from the bad.

Start by learning to use the "Search" function. Don't forget that many people incorrectly spell "castor" as "caster". :) Also "centre" as "center", etc., so use all the different spellings. And "offset" is often called "scrub radius", "trail" might be called "castor offset", and so on. Remember, this is a cottage industry, so nothing is done with very much rigour.
~o0o~

But next comes the really hard part. How do you decide which advice to believe, and which to ignore?

For example, back on page 1 Bill says,
"Because tire's Mz peaks out before Fy, does, a driver reading Mz induced steering force may fail to sense and hence operate the race car at it's maximum lateral acceleration. To reduce this enigma, caster is added to shift the peak Mz slip angle gradient upward..." (my emphasis).

I would say that Castor, IN ITSELF (!), has negligible influence on Mz. Instead, I would add Trail to the steering-geometry whenever I want to get better steering feel at high lateral Gs.

But the big question is; Do you believe Bill Cobb, who is an acknowledged expert and veteran of the industry, and gives his full name, or do you believe Z, who apparently has no track record whatsoever, and has a name that could only belong to a 13 year-old Mongolian girl who is here because she is bored tending to her yaks? :)

You might try resolving this dilemma by asking Bill for some sort of rational explanation for why things will work the way he has advised. I would suggest that Bill was thinking about typical production cars, where changing the Castor-angle of the upright ALSO changes the Trail. Again, IMO, it is this "side-effect" of the increasing Trail that gives more Mz "feel", but not the increasing Castor IN ITSELF (and I can give you a very detailed Mechanical explanation for why this is so, if you want).
~o0o~

As another example, Pat Clarke referred you to some advice he gave to new teams back in 2005. There, amongst other things, Pat advises FSAEers that;
1. Because of the slow speeds in FSAE there is "... little aero download assistance available ..."
2. "A mobile roll axis will send confusing feedback to the driver, making accurate control difficult..."
3. "Push rod or pull rod suspension is a good idea for the following reasons. ... Unsprung weight may be decreased "

Since 2005, the above three points have been argued against many times, and, again IMO, shown to be wrong by rational, detailed, Mechanical analyses (these are scattered throughout this Forum).

However, even though Pat repeatedly tells students that "... it does not matter whether the Judges agree with your design decisions or not, as long as you can cogently argue in favour of your solution." he has consistently refused to provide any "cogent arguments" supporting his 3 points above, (or many others).
~o0o~

So, Luis, it is up to you to decide who you believe, and why. Your choice is essentially between,
1. Dogmatic and unjustified opinions of experts who tell you to believe them because they have high "reputations".
2. Rational explanations, preferably with reliable empirical and theoretical support, that can be delivered by anyone (including 13 year-old Mongolian yak-herding girls), but that are nevertheless based on "reason".

So, "REPUTATION", or "REASON"??? :)
~o0o~

Oh, and BTW here are some of my opinions related to your original questions, from the "Caster/Camber/Steering shims" thread (http://www.fsae.com/forums/showthread.php?5332-Caster-Camber-Steering-shims&p=92714&viewfull=1#post92714) of 2005. I can support these opinions with as much detailed analysis as you want. It is only a short thread, so it is worthwhile for you to read it all, for the other opinions.

You might also want to read the "Reasoning your way through the FSAE Design process" thread for an overview of a good approach to the whole event (it is "stickied" at top of page). And possibly also the "Suspension Design" thread, although here you will mostly see how LITTLE reason is often applied to the design process.

Enough for now.

Z

[Kevin Hayward]

Luis,

Also have a look at the camber change due to steering from a KPI angle. Typically we are told that KPI is the devil's angle, but I think it is anything but. For example have a think of the camber on the inside tyre at initiation of turn. At the point it still has some load and is contributing to initial turn in.

I would listen to Z when it comes to considering trails. Both the scrub raidus and mechanical trail have a big influence on feedback and forces. One of the simple studies you can do is:

- Setup the 3d statics calculations for a double a-arm (or other) suspension system. This is easy to do in excel and is just a lot of simple calculations that you would have learnt in engineering mechanics (statics)
- Assume load is passing through the nominal centre of the tyre. (Note that deflection can change this significantly)
- Sweep the tyre load around a nominal g-g circle (i.e try reasonable combinations of vertical, lateral and longitudinal)
- Change castor's / trails / KPI's etc and look at the changes in the link forces

This will teach you a lot. Remember that the driver ultimately resists the tie-rod loads in the front end of the car. This simple spreadsheet becomes incedibly useful when you want to look at the mechanical design of the actual components.

Also try to keep in mind the basic function of a suspension system is to control the forces (e.g. tyre vertical load variation, tie rod forces, link forces) and the position of the tyres (including angles). All of the various suspension systems invented are attempts to do this in a way that maximises performance and/or reduce cost. There is no ideal suspension system for all circumstances. A double a-arm setup with pushrods, anti-roll bars, 4 way adjustable shocks, pitch spring on both ends of the car might not be the best system for you to implement. For example did you know one of the best places for a new team to gain points in a competition is in the cost report? If you assume a lower dynamic performance compared to the top runners, the final cost of your car ends up being a higher percentage of the points you will end up scoring. Maybe a beginning team might be best served to make different decisions to the top teams, to maximise competition performance.

Also note that it takes a lot of effort ot gain understanding in this area, and there doesn't seem to be any end to what you can learn. At some point you need to jump in and do the calculations from basic static and dynamic principles yourself. The sooner you start the sooner you will escape being misled by the plethora of bad advice out there ... including my own.

Kev

[Jonny Rochester]

In addiction my car have 0.5 degree of toe variation dynamically on the roll, and again I don't know if it is a bad thing.

Which way does the wheel toe in? When the front suspension is compressed, do the front wheels toe-in or toe-out? If they toe-in, that will be bad (making the car unstable). If the wheels toe out, this change maybe be acceptable, or make the car more stable.

[LuisN]

Z,

The journey that I mentioned is the process of designing my car, that is in the middle. But the journey that is learn vehicle dynamics I'm near the start.

Thanks to point out that I need to reason more advice that I was given.

Now the mechanical part.

Correct me if I'm wrong. By adding trail to the car, I induce more self-align torque so the feedback in the wheel for the driver is bigger. In addiction I can do this by making the arm in Z, on the steering arm, related which the center of the wheel bigger that cause the same effect of feedback, but reduce the self-align torque felt by the tire.

Thanks

-o-O-o-

Kev,

I started building one excel sheet to calculate all forces in the car suspension. For doing this I decided to convert all links in vectors, then apply the forces.

There is a better way? Or simple way? (Making the static equations by hand then solving they in a matrix on excel? My aim doing this sheet is one sheet for all cars after me, that changing some suspension data you are able to get all forces on it)

Furthermore, I didn't realized why KPI is the devil angle?

Thanks

[Claude Rouelle]

LuisN,

Here are some perspectives which could help you to look at the large picture.

If you want your car to get as fast as possible there are a few conditions
- Get the best possible tires
- Get to use each of the tires at the maximum of their traction ellipse (combined Fy and Fx) and all times whenever possible. I say whenever possible because sometimes the limitation will be other factors. As an example your engine power, more than your tire longitudinal grip, could be the limiting factor of your acceleration.
- Get the yaw moment you want when you want. There are 13 causes for the yaw moment: 12 mechanical: 4 tires Fy, 4 tires Fx, 4 tires Mz and and one aerodynamical: the aerodynamics Mz on the whole car (Aero Mz is not negligible even at low speed, if you have big aero devices such as wings and big wing endplates)
- Get the car driveable (difficult topic which mixed engineering objectives analysis and subjectives data: what is driveable for driver A is not necessarily driveable for driver B )

Each specific tire will give you on a specific asphalt a given Fx and Fy and Mz (and/or a combination of those) and also a Mx (but lest not go to far at this stage) depending the rim width and shape, the vertical load, the camber, the slip angle, the slip ratio, the tire pressure, the speed, the tire wear and the tire temperature.

Coming back to your specific question and if you essentially look at camber and vertical load effect on tire lateral Fy and longitudinal Fx grip and Mz you need to understand the effect that KPI and Caster angles and trails will have on:
1. Each tire loaded radius change (or wheel center height change if you want)
2. The camber variation is steering
3. The vertical load change on each tire in steering. I often ask students " Lets' say your car in on scales in your workshop and you turn the steering wheel to the left (so the car has no speed and no lateral G and no lateral or longitudinal weight transfer) is the RF corner weight increasing, decreasing of staying the same and can you explain why?" and 2/3 of the time I have the wrong answer. If you can explain that then I can tell you that you already understand a big part of what KPI and Caster angles and trails do.
4. The ride height change (some call it the steering jacking) and there fore the pitch angle change
5. The steering torque in steering
6. One you understand this in steady state you can evaluate how much in transient the SPEED at which you turn the steering wheel will change the SPEED at which each tire vertical load will change.

Now I can give you a course that describes step by step the influence of Caster and KPI angles and trails on the above parameters but that would be too long to describe in a forum. I teach this material in seminar (there will be one on in Brazil beginning of November) but there are already very good perspectives and information that you can find on this topic in excellent books and in this forum.

I hope this helps. I suggest you inform yourself more by searching and reading books/articles/forums and make your own basic calculation in Excel and then if you need you need/want, you come back with more specific questions; I (and others, I am sure) will be available for answers.

[Kevin Hayward]

LuisN,

Sounds like you are on the right track, with converting all the links to vectors and solving with a matrix in excel (or matlab). I won't go over the statements about KPI, other than to say keep reading and you will find the comments. The point was that it has an effect on camber variation while steering, and will affect your scrub radius, and you don't always have to have it at 0 degrees.

Also if you get a chance to go to one of Claude's seminars do it. It is well worth the money, and while no one I have met has a complete understanding of vehicle dynamics, Claude is brave enough to organise his understanding and get it out there. I went to his 3 day seminar quite a while ago, after I had done years of study and reading. It really helped direct my thinking and got me pursuing new directions. My only complaint was that the days were far too short at only 11-12 hours each day.

Kev

[Claude Rouelle]

Kevin,

Thank you for your nice and encouraging words about the OptimumG seminar.

In fact you should come back and attend it once more; you would be amazed how much it changed in content, quality, quantity, and usefulness.

To address the issues of long hours and the huge amount of information per day the seminar is now 4 day on theory and we also give a few times a year another 3 or 4 day seminar with practical training in smaller groups with a lot of simulation and data analysis exercise (you need to bring your laptop).

But the biggest difference compared to the seminar you participated in is that the videos I show are funnier and the jokes I tell are ... worse :)

[LuisN]

Claude Rouelle,

Thanks for the brief explanation. Your baggage in Vehicle Dynamics is my goal to achieve one day.

Thinking on that I applied to your seminar here, it is just a pity being after the competition.

And Kevin thanks for the positive feedback of the course.

Thanks all.