Suspension Design

I am currently working on designing the suspension of a powerful offroader, i decided to start from scratch and have a look at most suspension types.

I was very impressed with the sliding pillar for such an application, it is kind of like an active suspension but more simple. i decided however to against it because of:

-for me and where i am located manufacturing and assembling its components will be quite difficult
- roll angle influence on camber, i can't use a too stiff anti roll bar as with an offroader (susceptible to many single wheel bumps) it will greatly influence ride comfort.
- to design correctly will take a lot of time which i don't really have.

I would say this type of suspension is quite ideal for long term development and the potential gains from it are there to be found.

In the end i went with a configuration of double wishbones in the front and a De-dion in the rear (which does have its negatives)

for Z:

you were previously saying a De-dion is suitable for FSAE, while it is quite good and has benefits but i don't agree and to a limit i also don't agree with the sliding pillar. all vs using double wishbone.

- with a De-dion you should have superior traction but why? because the wheels will be perpendicular? that can be arranged with a SLA as well so what is the advantage here?
-simplicity, maybe in design but not in cost, weight or packaging. you will save some weight on the upright and some with the links and instead will replace them with a very long beam that could be just as heavy + watt linkage and other control links, and will need extra space behind the suspension. teams will mostly use rockers to activate the spring and damper and perhaps attach an ARB (why not use the de-dion as ARB?) for tuneability so not much saving when coming to links.
- Given the above the only thing a De-dion will save on is perhaps rod ends costs, and that won't be much really and even them most teams get them for free

Sliding Pillar:

-Potential wise it is a very impressive system as i said above.
- the design of a successfully operating system is not easy, and i believe should be much harder than a SLA system.
- Tolerances and fits will be critical
- the system is not popular and thus trouble shooting won't be easy.
- with this complexity in design and manufacturing, it will most certainly need more time to be made, for sure more than the already known SLA setup.
- I think it will be more expensive than SLA as well.


Theoretically, a lot of suspension systems can work on an FSAE car, but you are overlooking an important fact. these teams are amateurs who have very limited time, resources and need something adjustable (for testing on rough roads and sponsor days) and tuneable so that when they screw up they can adjust themselves and give their drivers decent practice time.


the SLA's beauty is not that it is the one which is used frequently (and i do admit that most will use it simply because it is the most common) but more that even though it's design process and optimization is not easy, the beauty of this system is that it will work no matter how bad you screw it up, and even if the team screws it up beyond repair they will just brake the Arms and make something new on the spot with no design and see how it goes. with most other systems the packaging process should be very neat and major problems will have to wait till next year to be fixed, oh and if they do fail they will fail miserably and repair won't be easy.


Please do correct me if i am wrong.

Axel,

"... at 3-10 hz, you're basically just saying use rumble strips instead of cones."

Well, longish wavelength rumble strips that are the full width of the track, with cones either side so the cars MUST drive over the bumps.

Imagine a large radius, sweeping bend with many bumps/corrugations about 2 to 5 cm high, with 2 to 5 metre wavelength (like the roads around here... http://fsae.com/groupee_common/emoticons/icon_smile.gif ). A stiffly sprung car will have its wheels lift off the ground at relatively low speed, and will thus lose grip and drift outwards, knocking over cones. So it has to slow down. A well suspended car (ie. softly sprung and correctly damped) takes the corner at whatever Gs its tyres would give on a flat course. So it wins. http://fsae.com/groupee_common/emoticons/icon_cool.gif

BTW, that yellow car (Concordia) has 8" rims, so the sliding pillars couldn't quite fit inside the wheels. If you look closely, the suspension is "outboard" of the steering, so no bump-steer.
~~~~~~~~~~o0o~~~~~~~~~~

M. Nader,

I will (briefly) cover what I think is the most suitable suspension type for off-road in a later post.
~o0o~

Regarding De-Dions:

"- with a De-dion you should have superior traction but why? because the wheels will be perpendicular? that can be arranged with a SLA as well so what is the advantage here?"

With ANY independent suspension you always have to compromise on the wheel inclination angle that you get with roll, versus that with pitch/heave of the body.

So, with 0% camber compensation (= very long Front-View-Swing-Arm) the wheels remain perpendicular to the road during pitch and heave (eg. accelerating and braking), but lean outwards during cornering to the same extent as the body-roll. With 100% camber compensation (FVSA = half track) the wheels remain perpendicular during cornering (regardless of amount of body roll), but lean in or out during body pitch and heave.

So with independent suspensions you either have to stiffen the roll-mode (with ARBs = bad), or stiffen the pitch/heave-modes (with "third-springs" = less bad). With beams you can have all-round soft springs, and no problems with adverse wheel camber.

Over the years there have been many attempts to solve the camber problem with independent suspensions (eg. the Dax thread), but really all they do is reconnect the "independent' suspensions so they behave like a beam-axle (but with a much more complicated linkage!). There are some more comments on beam-axles, etc., on the Dax thread.
~o0o~

"-simplicity, maybe in design but not in cost, weight or packaging."

Oz Olly and Chapo (ADFA team) reckoned the whole build process was a lot easier with their twin-beam car, than with their earlier wishbone cars (see earlier posts this thread).

A key point that they mentioned, and I will restress, is THE CHASSIS BECOMES MUCH SIMPLER. There are far fewer chassis "hardpoints" required with beams. The sketch I posted on the Beam-Axle thread only requires 4 hardpoints per beam. When I get around to posting a De-Dion sketch on that thread (gotta buy right computer cables first, grrrr...), it only requires 3 hardpoints.
~o0o~

"Sliding Pillar:
-Potential wise it is a very impressive system as i said above.
... [harder design, tolerances are critical, more time...] ...
- I think it will be more expensive than SLA as well."

I agree with all your comments there. As I said, it is not my favourite. But perhaps for a team like Amberg-Weiden, who use wheel bearings that probably cost more than many other FSAEer's whole cars... http://fsae.com/groupee_common/emoticons/icon_smile.gif
~o0o~

"... these teams are amateurs who have very limited time, resources and need something adjustable ... so that when they screw up they can adjust ...
... the beauty of [double-wishbones] is that it will work no matter how bad you screw it up..."

Well, not really.

I have seen a great many amateur, albeit very wealthy, Gentleman racers, who invariably "screw it up" so bad that the final adjustment that they have to make is to fit springs so stiff that the suspension "doesn't work" (ie. "Any suspension will work, if you don't let it.").

The main problem is all the adjustability. This is a great toy that the amateurs love playing with, but more often than not it leads them to places that don't work at all well (rear bump toe-out being a common "bad place"). So the universal solution is "lock her up!".

A much simpler suspension, with just some approximately vertical up-down movement of the wheelprint, minimal (perhaps unadjustable) bump-steer, and some adjustability of static-toe and camber, and spring rates, is more likely to lead to a suspension setup that actually works (ie. it moves!, it soaks up the bumps, and the car is fast).

Z

"With ANY independent suspension you always have to compromise on the wheel inclination angle that you get with roll, versus that with pitch/heave of the body."

I agree with that 100% in fact the equations clearly say so, but with a SLA you can do the adjustments in many ways using Kinematics and/or anti-roll bars (which are not that bad for FSAE considering the lack of single wheel bumps). so still the de-dion s not that much better than SLA.

+ the De-dion will also pose the single wheel bump problem, and camber change on the other wheel.

for me the De-dion i am designing will have 6 pickups. 2 for the rocker/damper (i MUST have them inboard, or adjustable to a 15cm length!) and 1 for the Watt on each side. 3+3.

A much simpler suspension, with just some approximately vertical up-down movement of the wheelprint, minimal (perhaps unadjustable) bump-steer, and some adjustability of static-toe and camber, and spring rates, is more likely to lead to a suspension setup that actually works (ie. it moves!, it soaks up the bumps, and the car is fast).

but isn't that too adjustable?, i mean this is basically what every team is aiming for and that is why they use the SLA, for ease of adjust-ability of such parameters. I think no teams want to vary track length or pickup points, but only focus on toe, camber and ARB. (+ perhaps some Motion ratios for the wheel rates). and would you agree that a bunch of amateurs playing around with a SLA is potentially safer (for the car's performance) than them playing around with a live axle for example?


I think you know now that i am not bashing down any of these systems, but in no way are they superior to a well designed SLA setup as an overall package, as each of them will have his merits and faults. i think for FSAE the team should choose what "they think" is the easiest to make and which works with their tuning goals in the end. It won't really matter as they can make any of these setups work.

But yes, it is very much worthwhile to open the eyes of some that there are other systems out there other than the FSAE standard SLA (i didn't have that when starting the project) as new entrants will only look to the SLA and not even know what is their because they will think they are in beyond themselves so will just choose what others use. while we are at it, why not use the 4 bar system in RCVD as well, it kinda takes the best of both worlds!

Now hurry up with that offroad suspensions, i am looking forward to what you can bring to the table on that one. http://fsae.com/groupee_common/emoti...on_biggrin.gif

Z,

That Dax thread is from 7 years ago...
I had no idea you had been pushing the beam axle concept for so long!
Your persistence does you credit.

-William

Quote:

---

Originally posted by Z:
But perhaps for a team like Amberg-Weiden, who use wheel bearings that probably cost more than many other FSAEer's whole cars... http://fsae.com/groupee_common/emoticons/icon_smile.gif

---

Small correction: Amberg-Weidens wheel bearings are probably cheaper than any of the top teams and at the same level as most of the mid-class teams ball bearings.

Their solution for bearings is actually very clever but I don't know how much compliance / friction it has.

Quote:

---

Originally posted by Markus:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Z:
But perhaps for a team like Amberg-Weiden, who use wheel bearings that probably cost more than many other FSAEer's whole cars... http://fsae.com/groupee_common/emoticons/icon_smile.gif

---

Small correction: Amberg-Weidens wheel bearings are probably cheaper than any of the top teams and at the same level as most of the mid-class teams ball bearings.

Their solution for bearings is actually very clever but I don't know how much compliance / friction it has. </div></BLOCKQUOTE>

didn't they say it was 30% more expensive than the wheel/hub/spindle/tapered bearings combo?

Something along those lines but I think that might be the cost event price for the corner i.e. != real price.

I would say the hub / upright is the expensive part in their design, not the bearing.

M. Nader,

"... use the SLA, for ease of adjust-ability of such parameters [as] toe, camber and ARB. (+ perhaps some Motion ratios for the wheel rates)"

The thing is, "toe and camber" are two of the main adjustments you use when you have NO suspension at all! Throw in tyre pressures, steering geometry, and chassis stiffness, and you can control LLTD and pretty much all of handling balance, still with NO SUSPENSION.

"... would you agree that a bunch of amateurs playing around with a SLA is potentially safer (for the car's performance) than them playing around with a live axle for example?"

No. Again, the problem is the SLA has far more "adjustments" that can be made, and the amateurs (and a great many professionals) get lost.

This applies during design where a huge amount of time is wasted (What RC heights? RC migrations? Pitch centres/migrations? Long or short VSALs? Toe-changes? Rising or falling spring-rates???). And then again when setting the car-up, because the designers cleverly (!) built-in a lot of optional chassis attachment points, "just in case".

"So, if we move these brackets to here, ummm, ... that'll give us more anti-dive..., which should cure the excessive neg-camber during braking... err..., but then what happens to our RCs? Ummmn... and the toe-curves...???"

Like I said, they make great toys...
~~~o0o~~~

"Now hurry up with that offroad suspensions, i am looking forward to what you can bring to the table on that one. http://fsae.com/groupee_common/emoti...on_biggrin.gif "

Well, this could fill a book, but briefly (and roughly in order of importance).

REQUIREMENTS of OFF-ROAD RACING SUSPENSIONS.
=================================================

1. Huge vertical wheel travel.
This trumps just about all else. It is why front-engine+rear-live-axle "truggies" (truck/buggies) are still competitive off-road. The combination of a longish (~2 m) prop-shaft plus roll motion of the axle means that one metre wheel travel (or more) is possible to the driven wheels using only two off-the-shelf UJs. Even with very wide tracks the rear-engine+transaxle cars only get about 0.6m travel with conventional CV'd halfshafts. (There is a simple solution for 1+m rear independent suspension, though... http://fsae.com/groupee_common/emoticons/icon_wink.gif )

2. Relatively soft springing.
The above travel is useless if not used. Best is to have a very soft twist-mode. The truggies typically have axle-roll softer than axle-bounce (by mounting springs closer together on beam), which helps soften the twist-mode. Rock-Crawlers also do this, with live beams front and rear, each with highish RC and soft roll springing. ARBs stiffen the twist-mode as much as they stiffen roll-mode, so they are almost never seen off-road (I have NEVER seen them, but...?). One day off-roaders might rediscover interconnected suspension (like 1930s 2CV design), but I'm not holding my breath...

3. No wheel camber change in bounce.
Heavy, large diameter, and rapidly spinning wheels exert huge gyroscopic couples when forced to change their direction of rotation (calc the numbers, they are big!). So Pure-Trailing-Arms at rear are good (and very common). VW Beetle style Twin-Trailing-Arms (a type of double-wishbone) at front are also good (but put a lot of mass out front, and hit banks because of bad "entry" angle). Currently fashionable Lateral-Double-Wishbones are (almost?) invariably equal length and parallel to give zero camber change. Unfortunately, these give hugely varying "jacking" forces in corners, and many drivers hate them, but still use them "because that's what real racecars have...".

4. No, or small, lateral wheelprint movement (= "scrub") in bounce.
This means close to horizontal front-view n-lines (ie. ground level RC, wheelprint moves vertically). Any lateral wheelprint movement over bumps causes lateral forces between wheelprint and body, hence either breaks tyre grip, or shakes car, or a bit of both. Also, lots of lateral movement (= high RC, = steep n-lines) implies lots of jacking forces for independent suspension, so lots of vertical motion of body when using soft springs, and greater possibility of rollover.

5. Wheels move backwards in bounce.
This lessens forces as wheels climb over bumps (note, this also has a lot to do with changing rotational speed of the wheel, rather than just the rearward road-to-wheel forces). This wheel "recession" comes automatically with rear Trailing-Arms, and Beetle-style front TTAs. With the Lateral-DWs the wishbone chassis mounts are tilted up-at-front to allow this backwards movement. This gives significant pro-dive under braking, but real off-roaders don't brake! Well, the front brakes are tiny... http://fsae.com/groupee_common/emoticons/icon_smile.gif

Oh..., and everything is BIG, and STRONG! Definitely, NO FSAE 5mm BJs!!! http://fsae.com/groupee_common/emoticons/icon_smile.gif
~~~~~~~~~~o0o~~~~~~~~~~

Will,

"That Dax thread is from 7 years ago...
I had no idea you had been pushing the beam axle concept for so long!"

I should point out that FSAE is one of the few places I would use beams. I, personally, would NOT use them for a top-speed off-roader (maybe for Trials or Rock-Crawling), and they take up too much room for passenger vehicles (ie. bad packaging). However, for FSAE they are a very good "fit" (quick and easy design and build), especially if you want to do aero-direct-to-wheels.

The main reason I keep pushing beams here is that I know that quite a few students agree with me, and would like to go that way, but they are dissuaded by completely irrational (and thoughtless, moronic, etc., etc...) arguments from their peers (bottom of Brent Butler post) and the judges (bottom of Clausen post).

I have no problem with people wanting to follow the flock, but I don't think it at all good for society when the braindead insist that everyone else do as they do.

Much of this thread shows how obstinately some people resist change, and how fiercely they condemn others who want to try something different. And, for some reason, in this fight to maintain the status quo, never is a rational argument used... http://fsae.com/groupee_common/emoti...n_confused.gif

It's just not engineering! http://fsae.com/groupee_common/emoticons/icon_smile.gif

Z

With regards to off roaders:

1. Huge vertical wheel travel.

Yup agree, but the exact value of that is dictated by the roads used, if used on soft or moderate desert roads you won't need more than 0.3-0.4m of travel. for huge offroad cliff jumps or mountain riding that number can go up to 0.6-1m.

2. Relatively soft springing.

Yes again, i would think a ride frequency around the 1Hz range (more or less) would be sufficient?

3. No wheel camber change in bounce.

Here is one the problems. for an offroader the aim will always be less camber change in bumps, but then when turning ( i wouldn't think a normal/amateur will be able to get more that 0.4G in the desert roads) and with the relatively soft roll gradient required for driver comfort (nothing less than 5deg/g?) the camber gain would be around 2 degrees and that would yield substantial loss of traction on both ends. so i wouldn't say full parallel A-arms are a good solution, maybe a a wide FVSAL but with a small consideration for roll would be better with a static camber of around 1 degree.

4. No, or small, lateral wheelprint movement (= "scrub") in bounce. This means close to horizontal front-view n-lines (ie. ground level RC, wheelprint moves vertically)

A ground level RC/IC is nearly impossible to get in offroaders, the chassis is quite a distance off the ground (let's say 0.5m) so basically you will need a 20 inch wheel so that the upper wishbone can just be at the same level as the lowest point in the chassis, I don't think this will change with any independent suspension, only independent ones will get a horizontal IC/RC close to the ground but will still be around 0.4-0.5 higher. with any independent suspension the IC will be pretty high since the lower wishbone will have a very steep upward direction and so would the upper one (for less camber variation in bump+ limited by rim diameter), resulting in a high and wide IC, resulting in a high RC/n line slopes. in general i would say the RC/n-lines/FAP will be higher than the CG, but the designer should make it as close as possible to have less rolling torque and thus less roll and camber variation in roll.


5. Wheels move backwards in bounce

I never really thought of that, would you care to explain why this would lessen the forces? wheelbase variations will be very small in either case and i didn't really think they will do much to reduce forces sine the force paths through the wishbones and n-lines won't really change significantly as the contact patch's position changes.


Trailing arms + (semi)dependent rear suspensions i would say have a clear advantage in the rear of offroaders (a consideration for drive shaft length change should also be made here if it is rear wheel driver), the front will depend on many factors and really most independent systems should work well there.
ARB most negative side in offroaders would be single wheel bumps, so either no ARB should be used or a very soft one with contribution no more than 10-20% of springs anti roll torque just for driver adjust-ability (most won't really need this, as time can be gained elsewhere).


and yes i know, FSAE sizes/suspension won't really work when a heavy (a single seat offroader with a strong engine i would say weigh around 300-450 kg to say the least) car is jumping 1 meter off the ground http://fsae.com/groupee_common/emoticons/icon_smile.gif

The wheels moving backwards in bounce reduces loading because you are reducing the vertical acceleration required of the assembly as it goes over a bump. Think about it: if your wheel hits a bump and the geometry tells it to move forward as it does so, you are causing the wheel to accelerate up the bump at a greater rate than if it moved backwards. Hard to explain, relatively easy to picture.