'Optimized' Camber Curve

[Chris Razl]

I'm working on the suspension design and I'm somewhat troubled because I cant find good information regarding camber gain. It is cleary a trade off between performance in a rolled state, and performance in braking and accelleration.

It seems to me that if roll is all you're worried about, you would want your camber gain to be equal to the chassis roll so that your contact patch stays flat. This is what most books and references seem to suggest is ideal.

However, you would then get large camber change under braking and on the gas, reducing traction. Of course with no camber gain, you maximize braking and accel, but you don't corner as well.

It also seems that (having played with the numbers and eyeballing other cars, fsae and not) most designs do not have camber curves as aggressive as this 1:1 ratio camberchange:chassisroll, but have some intermediate ratio.

It is simple to model the geometry and kinematics of wheel motion, but difficult to predict the result in lap times. Where, between the two extremes, is the 'fastest' compromise for FSAE? How can it be modelled and verified?

Chris Razl
Queen's Formula SAE

[Denny Trimble]

Will your car gain more in lap time from cornering performance, braking performance, or traction-limited acceleration performance? Clearly cornering, as it's where the car spends most of its time.

As a rule of thumb, bias your camber curves toward cornering, but don't go as far as 1/2 track FVSA lengths.

Also look into caster for additional helpful camber up front in the tighter corners.

And, beam axles give you perfect camber curves...

[Chris Razl]

Thanks for the reply, but I'm interested more in a logical, scientific method of determining what would be the best geometry. I mean, I have an intuitive idea of where it should be, but...I'm to tell that to a design judge?

I've been to comp twice and would say most teams run less camber gain than that needed to offset chassis roll. The cars that are the exception to the rule have suspension geometries that look like this, which looks exactly like what my numbers are spitting out:

The lovely UWA car:
http://uwafsae.ee.uwa.edu.au/downloads/Pics/IMG_2444.jpg

As for your comment...the car above uses a 1/2 track long arm (which you explicitly told me to avoid)...it has much more aggressive camber gain than the fsae norm...but wait...they scored 2nd on skidpad, 2nd in endurance, and 12th in autocross...

isn't that kinda conspicuous?
Chris

[mtg]

Use tire data (that you bought thanks to Denny, or measure yourself) to calculate tire force vs lateral and longitudinal load with different camber angles.

Once you do that, you will have some scientific and objective data to pick a camber curve.

But that's not the whole story http://fsae.com/groupee_common/emoticons/icon_smile.gif There's a million other things at work. How does the rate of camber change affect the tire, how do the camber curves affect the roll centers.....

The explanation of UWA's virtual swingarms to results is a fallacy. You could have terrible suspension geometry and win competition if you had everything else in line. That's not to say UWA has undesireable geometry- they do seem to know a thing or two about mechanical grip. When working with Kevin, you tend to get that impression.

So, you're looking for the OptimumCamber? http://fsae.com/groupee_common/emoticons/icon_smile.gif

[James Waltman]

UWA had some other magic going on with their suspension.

I wouldn't put much weight on competition ranking vs. geometry. University of Wisconsin-Madison had very different geometry than most in 2004. Still, they were 1st in Skid Pad, 7th in endurance, and 12th in Acceleration.
http://dot.etec.wwu.edu/fsae/James/Detroit%202004/Other%20Teams/U%20Wisc%20Mad%20rearsmall.jpg


Buying the FSAE Tire Test Consortium data would be a good place to start. More details here. (http://www.millikenresearch.com/fsaettc.html)

[Frank]

ditto denny, except the beam axle bit

ill need a bong to even think about a beam axle on our car

[Kevin Hayward]

Chris,

UWA's Swing arm lengths were not 1/2 track.

The rear was 1200mm (little over full track at the rear) and the front was 2300mm. The front of the car being much more dependant on castor angle than on swing arm length. I think these numbers are pretty much on par with a lot of teams.

I think the current plan is to extend the swing arms. When you increase roll resistance you can get away with a little less roll compensation.

If you want a scientific way to set camber curves it needs to be done on the back of testing. Attach some infra-red tyre temp sensors to your car and go running. You will find that the type of tyre matters. The sidewall stiffness is very important. One example from our testing is that the Goodyear tyres are very wear sensitive to camber even if they are a little faster. The thing to learn from that is to take the tyre test data with a grain of salt. Nothing beats on-track testing.

Analysing the data to find out what you should be running is a whole new topic. A really simple way is to break your data down into corners that represent low, medium and high speed corners and look at your temperature distribution in each of them. Remember to keep a reference of your steering angle for the front as it will have a great affect for the castor angles most of these cars run. You should be able to differentiate between the effect of static camber and dynamic camber.

I would not over-estimate the effect of geometry on a skidpan. Most of our time gains on the skidpan were made through damper work and driver training. Have a look at the data of a skidpan run and notice how much lateral g's are dropped for just a small steering or throttle change. Also note how difficult it is to hold constant throttle and steering on a really bumpy track.

When it comes down to it if you do not have the data (which Denny certainly does) you cannot make a better choice than some sort of subjective weighing up of the factors. I don't know about other people on this forum but my intuition makes up for a heck of a lot of my ignorance.

Cheers,

Kev

Optimum G

[Rob86]

Hi,

I dont know how many bongs Adelaide were smoking but they came first on the skidpan last year, above UWA, using a rear beam axle.


Also look into caster for additional helpful camber up front in the tighter corners.



To be more specific should this be KPI? Your caster could be all mechanical trail and give no camber gain for steering angle.

[Z]

Originally posted by Chris Razl:
I'm interested more in a logical, scientific method of determining what would be the best geometry.
Chris,

If you want to know what sort of camber curve is best, then first you must know what sort of spring stiffnesses you have!

Specifically, you must know how stiff your springing is in roll mode (cornering) vs pitch mode (acc/braking).


EXAMPLE 1. You have very stiff roll-mode springing. In this situation you don't need any camber compensation geometry - no body roll, so no camber change (other than a bit from tyre squash). So you can use parallel motion suspension (equal length parallel wishbones, or trailing arms, etc.).

Note that stiff roll-mode can be done while still have soft springing in bounce, and/or pitch, and/or twist modes (use "Find" on this forum for more info re. these "modes"). Also note that providing a stiff roll-mode with ARB's is a bad choice (it stiffens the twist-mode)!


EXAMPLE 2. You have soft roll-mode springing with fairly stiff pitch-mode. Eg. at each axle you have stiff "monoshocks" (= axle-bounce-mode springs), and soft ARB's (= axle-roll-mode springs). Now you should run full camber compensation, ie. FVSA = half track, as in the UW-M car in James' post above.

This might work on an aero car with low body mounted wings because the stiff monoshocks give stiff bounce and pitch control of the body (ie. stable aero platform) while the softer ARB's absorb single wheel bumps and twist in the road. Note that it is probably much easier (and lighter, stiffer, stronger, cheaper) to build this as an old-fashioned low-pivot swing-arm, rather than as the double-wishbone in the pic above.


EXAMPLE 3. You have soft springs all round (ie. soft bounce, pitch, roll, and twist modes). Now if you run beam-axles your cambers don't change during ANY body motions, so are always fully compensated http://fsae.com/groupee_common/emoticons/icon_wink.gif.

With soft springs you can also have relatively soft low speed damping rates and still be "overdamped" in all modes (ie. damping >= 0.7 x critical damping). So the body is very stable (no wallowing) but the suspension is still soft enough to absorb bumps and cope with twist in the rode. (Note: with stiff twist-mode springing the "twist in the road" dictates handling balance, not you the designer or set-up engineer.)


Oh, yes...
EXAMPLE 4. You have some sort of "in-between" roll stiffness from conventional spring-at-each-corner + 2 ARB's set-up. Now you go for some sort of compromise on camber compensation, so that it is not too bad during nose dive on braking, or during pure cornering, or over bumpy sections. But, then again, it is not too good...

So you tell the Design judges "It's all about making the right compromises, choosing the right balance between stiff handling and the ability to absorb bumps, err..., juggling the parameters, playing the right tunes... waffle, waffle...". I've heard this exact argument from an F1 chief designer, so I reckon the judges should buy it http://fsae.com/groupee_common/emoticons/icon_smile.gif

Z

[Kevin Hayward]

Rob,

Sorry to burst your bubble but UWA won the Oz Skidpan in 2004. Second in the US Skidpan by 0.002 or something really small.

Cheers,

Kev

[Frank]

the event was a lottery anyway.

I'd never consider the results a good comparison because there was pile of sand in the middle of the path.

SAE-A should be shot for letting it happen.

And the sand on the autocross track (there was IMO very little effort spent trying to remove it) was dangerous.

Your local hillclimb club wouldn't race under such conditions.

[Mike Claffey]

Frank I know everyone has excuses, but the time UWA won with, was from our first session, and there was dust blowing everwhere. The second sessions later in the day teams were shaving alotta time out on the cleaner track, but we had some issues with our second run preventing it from being any good. =[

On a side note, do any of you victorian guys know if they filled in the inside of the track with tarmac?

[Frank]

"but we had some issues with our second run preventing it from being any good"

yeah (scoff) so did we, a huge damn pile of sand in the middle of the track

[Kevin Hayward]

Frank,

I agree with you about the sand. It was a real problem that should have been fixed. Having said that I noticed that most teams stuffed the second circle by pushing too hard into the sand the first time around. Everyone knew the sand was there. Everyone had to deal with it. Once again no event is free from requiring driver skill.

The skidpan surfaces in Detroit are far worse than in Australia yet 0.1 seconds separates the top 8 competitors.

Even on our testing surface in Perth which is billiard table smooth the skidpan is never a steady state event. So what is the big problem with comparing cars over the same track that does not have the same grip the whole way around?

The same can be said about the Autocross. Yes it was sandy and crappy but both RMIT and Woolongong had there cars set up well and the drivers drove exceptionally.

All of that being said this post has got way off topic. However it does show that there is a lot more to performance around a skidpan than good kinematics.

Cheers,

Kev

p.s. Frank are you still with UQ or are you working now?

[Frank]

Kev,
Im @ UQ

Anyhow, camber curves, short VSA, that's what i reccon, not as short as half track length if the surface is bumpy..

put the pivot under the other wheel IMO

tyre data wont help, because the limit of Lateral G's (due to having a shorter VSA) is a function of the course smoothness

[Denny Trimble]

Originally posted by Rob86:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content"> Also look into caster for additional helpful camber up front in the tighter corners.


To be more specific should this be KPI? Your caster could be all mechanical trail and give no camber gain for steering angle. </div></BLOCKQUOTE>

Nope, caster angle (side view inclination of the steering axis from vertical). You can adjust your mechanical trail by offsetting the spindle fore/aft between the balljoints ("spindle offset"). It's possible to have 10 degrees of caster and 1/8" mechanical trail, if you wish.

[Rob86]

I stand corrected! Sincerest apologies...

One should go by facts and not hearsay

facts here (http://www.sae-a.com.au/fsae/results.htm)



Originally posted by Kevin Hayward:
Rob,

Sorry to burst your bubble but UWA won the Oz Skidpan in 2004. Second in the US Skidpan by 0.002 or something really small.

Cheers,

Kev



Nope, caster angle (side view inclination of the steering axis from vertical). You can adjust your mechanical trail by offsetting the spindle fore/aft between the balljoints ("spindle offset"). It's possible to have 10 degrees of caster and 1/8" mechanical trail, if you wish.

Isn't caster described in terms of a combination of pneumatic and mechanical trail?

[CMURacing - Prometheus]

caster is the angle the balljoints make with the vertical axis of your spindle. this, combined with spindle offset, determine mechanical trail. pneumatic trail is a property of the tire, and independent of geometry.

people (like our suspension lead) sometimes call mechanical trail "caster trail", which leads to confusing discussions.