--> Suspension Design <--

[Xeon]

Hey guys and gals our team is new to FS and we realise that Suspension design is pretty much key to the design of the car.

I was wondering if anyone has advice on how to go about calculating suspension mounting points, A arm lengths, angles, castor etc.

Also baring in mind we have no funding from our university at all during this point in time, so purchasing programs or apparatuses is not something we can do right now.

If anyone can post a Web link of decent but free suspension design program or post advice on how to go about manually designing that would be great.

[exFSAE]

In this series, reliability is key to the design of the car and trumps suspension design. Seen many "well designed" cars blow up in endurance and score few points.

In any event... from a true engineering standpoint, the first step is to establish the performance characteristics you want from your car. What's important to you? What performance is most critical on a FSAE-style circuit?

Buying 'Tune to Win' (Carroll Smith) is a good starting point for at least understanding some basics.

[woodsy96]

If you want software but need money, find money. A sponsor's bank account is a good start. Your's is a good backup.

Susprog3D and OptimumK both offer student discounts. Susprog3D is (from memory) the cheaper option.

[Auerbach]

2d sketches in CAD are a good place to start. You do have CAD, right? Also the 2d sketches can lead to point-based 3d sketches that will do 80% of what a suspension software program can do, it just takes more time and effort.

Also, I agree with exFSAE, reliability far outweighs suspension design, especially for a first year team. Also just packaging everything will be a big challenge if you've never designed a suspension.

[Xeon]

thanks, and yea i have Auto CAD 2010, solidworks 2010 and a fellow member has catia.

[erau_paul]

We designed our suspension this year using a combination of Race Car Vehicle Dynamics (for theory), CATIA (for 2D sketches), OptimumK (for kinematics), this forum (for advice), and last years spec sheet (for good starting points). A group of us were able to start from virtually no suspension knowledge to something our advisor thinks is pretty decent in about 2-3 weeks. Not bad in my opinion.

[exFSAE]

I'll add... there's no magic suspension design that's "optimal" and designs itself. Everything is a trade-off. Even trying to conjure up some a-arm arrangement that gives perfect camber compensation is pointless... as if you choose to be perfect in bump you are terrible in roll, and vice versa. Unavoidable.

Really I think the key is to define what you want at a top level, instead of dicking around with points and seeing what they do. Once you define the overall idea... the parts and points really define themselves and do the work for you.

[Kevin Hayward]

Microsoft Excel, or for the free version Open Office.

Solving Double A-Arm suspension is just a case of solving four bar linkages. Quite straight forward equations that can be found in a number of cases. The suspension can be quite simplified quite well into the 2-d front on view making it a very easy problem. FSAE cars (as well as many others) are unlikely to have significant amounts of anti-dive or squat.

The main disadvantages are that the 2d case will not allow you to calculate toe change during bump. This isn't a disaster if you follow the basic guidelines of steering and toe-arm locations.

The big advantage is a bit of playing around with a 2d sim gives you quite a deep understanding of the trade-offs you are making with particular setups.

I haven't done the 3d calcs in excel, but I am sure it can be done. However they are not too much more difficult than the 3d ones. Once again 4 bar linkage calculations. Early efforts were focussed on geometric solutions using matlab. A free version would be SciLab or Octavia (hope I got the names right).

Programs like OptimumK generally use numerical solutions. This means that it is very easy to add new suspension types into the programs as a series of connected linkages. However for the Double A-Arm case the geometrical solutions are quite well documented and reasonably easy to implement. Undoubtedly you will have a lecturer in your university who is well versed in mechanisms that you can go and have a talk to. this will be quite straight forward for them.

Don't be scared by trying to do these calcs by yourself. They are not as difficult as it seems.

If you have the cash then go for OptimumK. There have been some truly brilliant people involved with its creation http://fsae.com/groupee_common/emoticons/icon_smile.gif

Kev

[erau_paul]

Kevin, do you use the loop closure equation for your design? I figured this would be difficult because of the number of unknowns. I would imagine that a guess and check method centered around the basic guidelines founds in books like RCVD is the most effective way of designing a suspension.

It's like my Controls professor taught us. There is an analytical solution to PID controllers but if you give a child 3 dials and tell him to play with them he will find the solution faster.

[Kevin Hayward]

I'm afraid my first efforts were much more simple than that and pretty straight forward in excel. I was the child with three dials, probably always will be.

Looking at the 2d front on view of the car I just swept the lower arm through the expected range of travel (one row for each increment). For each row you only then have two triangles to solve for using simple cosine rules to find the remaining a-arm points. You actually end up with two possible solutions, but only one makes sense and is easy to check for. This gives you the position of the points throughout their rotation.

At that point you have very simple similar triangles and geometric analysis to find out very useful things like track change, camber change, kinematic IC movement. From there it is simple to build other four bar linkages for rockers and corners together. This sort of spreadsheet is very quick and dirty to create. Chuck in different numbers to see their effect, inputing SAL's and RC heights to calculate initial x-y coordinates. doesn't take very long at all to figure out the main trade-offs.

Following that used a very similar method in 3-d in matlab. Sweep through the double a-arms so you know all the relevant positions and then you can do things like combined corners during a track run through previously tabulated data. It is slower than a good numerical solver, and not as elegant as neater close-fomed solutions, but it is very easy to code and check as every step makes a lot of sense.

I still hold that most of what you need can be found in 2-d and the reward to effort ratio decreases dramatically with increased complexity. By no means is this method very good to use in commercial programs or general purpose kinematics studies, but the double a-arm is just such a simple geometrical problem. High school level geometry is all that is required.

I also would advocate in the very early days designing the car to meet your analytical capabilities. There is no rule that says that the rockers (if you use them) have to be high (or low) mounted on funny planes that are more difficult to create in whatever CAD package you use. Just run them on the side of the car in a 2d plane shared with the suspension. Very easy to calculate for and understand. As for speed go and ask UTA if they'll wheel out their 2000 car with dampers just that position (not sure how long they continued like that) and see whether the increased aero drag or slightly dated appearance slowed them down at all.

By the way erau_paul if you are using OptimumK try and make good use of the spreadsheet import and export. It was created to expand the programs capabilities by allowing easy integration with Excel. If you export one of your suspension files you will get a well formatted excel sheet. That sheet will import quite well back into the program after any changes that you make. Additionally as long as you keep the first page unmolested (apart from the numbers) you can put anything on the other pages without causing input errors. This means you can create a second sheet in the spreadsheet that calculates your x-y-z components from more easily modified parameters such as SAL, scrub radius, KPI etc. This means instead of either mindlessly changing xyz numbers, or going through the tedium of a change in Catia then transferring numbers over one by one you can do very simple changes and see a quick result after an import. For instance use you calculation sheet to calculate the new xyz points for a change of caster from 3 to 4 degrees and see the effect on kinematics when run through OptimumK. Once the sheet is setup this involves changing one number, importing the sheet into OptimumK, then pressing run and then view the results. Much quicker way to work.

Kev

[js10coastr]

Originally posted by Kevin Hayward:

If you have the cash then go for OptimumK. There have been some truly brilliant people involved with its creation http://fsae.com/groupee_common/emoticons/icon_smile.gif


Brilliant and inspirational! You have no idea how much I learned from the team; technically, professionally and personally.

[Demon Of Speed]

Since I was there two years ago and really enjoyed the help, here is what I did.

1) Do a static analysis of the vehicle, taking into account the weight of everything that will be on the vehicle (including driver). You will need the know the size and weight of everything going on the vehicle, and if this is your first vehicle you probably don't know this stuff, so make good guesses and ask people. This should get you your location of front and rear axles.

2) pick a track width, make sure your car is wide enough so going around a turn you don't tip over. usually around 3/4 of the wheel base (WB).

3) Choose which tires/rimes you are going with. This is based on forces into the control-arms, space, and moment of inertia of the tires/wheels.

4) Choose a minimum turning radius, should be based on what you think the tightest turn you will see is plus a little more for under-steer and whatnot. Choose a percent Ackerman, think about the speeds traveling and slip angle encounters on all the tires.

5) Design the largest upright that will fit inside of the rims (remember that you have control arms coming off from the upright that can not contact the rim/tire during turning and bump/rebound). Add in Kingpin angle, mechanical trail, and caster angle as you think is necessary for your vehicle. the upright will have to handle the forces of braking, drive-off, and cornering (use FEA).

6) use a 4-bar linkage analysis to come up with the suspension kinematics that are desired. You know the mounting points on the upright, and around where the points can attach to the frame (templet size and track width). Setting up the 4-bar linkage is kind of different since the input is not directly angular. The String model is a good way to start, and easy to visualize what is going on.

6b) Main goals of the 4-bar linkage are usually to reduce weight transfer, keep the tires flat on the ground under all driving conditions, and reduce tire scrub. Look at tire scrub and camber at bump, rebound, and body roll. Roll stiffness should be looked at to determine how much the vehicle will roll.

7) Design the control-arm. Since you have no money use cheap steel (1018). Do a static analysis (assuming two-force members make up the control-arms) to see what kind of loads will be put on the control-arms during braking, drive-off, and cornering. Then using the infinite fatigue life of steel make sure the control arm member can handle that stress. Add in the bending moment due to hitting a cone on the center of the member as well as the locating point of the pull/push/CoD mounting location. Remember for compression to check for buckling. after the members are sized design them with FEA.

7.b) Choose on rod-end/spherical bearing sizing based on manufacturer info and what you need/want for your design. Make sure that the bolt going through the rod-end/spherical bearing is large enough so it will not shear in use.

8) Choose an actuation system for the Coil-over dampeners. This will be either pull-rod, push-rod, or direct actuation. They all have there pros and cons choose what works for you (mainly a packaging and MR problem).

8b) If using push/pull rod actuation and doing the calculations by hand for the MR (or in excel) keep everything in a single plane. I had everything on my car's front suspension in one plane and it wasn't bad doing the math (all circles basically), however I had things in multiple planes (two statically, three kinematiclly) and the math was a nightmare (spheres and planes all using matrix algebra). Remember to make sure to have a dampener that can fit where you design it to. Also, make sure to do stress/bucking analysis on the push/pull rod.

8c) Try and have you motion ratio as large as possible (around 1:1) to reduce forces on the frame and to get the dampener in the velocity range that it needs. Also you want a linear or progressive MR, why calculations should tell you why (and what you want).

8d) Choose a spring rate that keeps the car off the ground well driving, and has the desired roll stiffness.

8f) If there are handling problem an anti-roll bar might be needed, I don't like to have one in the initial design of the vehicle, only as a way to change handling later on if needed (car over-steers or under-steers to much).

Remember to have things adjustable, specifically camber, toe.

I did all of this in Excel and MathCad (there were 100+ pages of calculations), and FEA was done with Pro-E mechanica. The only thing that a program is nice for is the 4-bar linkage (especially in 3d), motion ratio design, and seeing the effects of caster, mechanical trail, and kingpin angle on the vehicle well turning. However all of this can be done in excel with a crap load of work. You can buy a program to do this for around $500 (http://www.mitchellsoftware.com) or spend 500+hr (literally) writing the equations yourself.

[Xeon]

Thanks alot Kev, i will try to crack this issue and will post how i find your method.