Recently there have been changes in the landscape of chassis engineering, and I've wondered if "everything I know is wrong".

The biggest issue: Force Based Moment Centers.

Why should I (or you) care?

Our Front View swing arm geometry effects our camber change and roll stiffness. This information can be used to enable purposeful and predictable changes to setup for the purpose of increasing performance. If your rules or inclinations permit you to tinker.

In the old milieu you might use Mitchell to evaluate your geometric options, and think you were doing pretty good. Most of it made sense. Vertical height of the Geometric Roll Center location - check. Camber change in jounce and roll - check. Lateral location of the Geometric Roll Center - Uh, not sure, kinda fuzzy, move on.

Then one day some Frenchman is quoted in an article on WRC saying something about "keeping the car rolling about the roll centers". A couple of years pass and now everybody who want's to be reasonably well informed is scrambling to rebuild a decent understanding of how a chassis and it's four contact patches interact in response to the driver turning the steering wheel.

I've been out of school a long time, and while I've lost alot of facility in analysis, I've also lost alot of naive assumptions about problem solving.

Look at a head on picture of a turning car. The chassis has rolled to an angle - about some axis - presumably a dynamic instantaneous one. At each point of the rotation from rest the locations of the suspension pivots are displaced - giving the instantaneous values of dynamic cambers, cornerweights, and chassis location.

I think I remember that rotation in a plane about some point can be represented as the sum of a linear displacement and rotation about another point. I think we're most concerned with the vertical component of the linear displacement and the rotation.

Of course if you're only turning left on banked tracks then things are a little different.

I've read some of Bob Bolles writing in Circle Track over the last year or so, and some of it has made little sense to me. I bought his book last week and read over the relevant sections. Ah, now he makes sense.

It hadn't occured to me that you might use asymetric front end geometry, but it makes sense in the oval context. They can make use of static roll center location offset from the centerline. Road racers will bias their camber settings for track rotation and critical corners, but I doubt many if any have biased their geometry. Interesting.

But, more relevant here is the fact that it's all based on conventional geometric roll center calculation - he's not talking about the same thing as the other guys who are talking about "Force Based Moment Centers".

In thinking and reading about this stuff I constantly have to keep asking why I care, and the more I think about it the less I care unless I remind myself why I care.

Plenty of race car designers are on record as saying that they don't concern themselves with geometry much. Most of them make use of significant downforce and minimal travel though, so I don't automatically accept their opinions as relevant to tin tops. And of course they could be lying. They can do that.

Yes, it's true: "any suspension will work if you don't let it". Ha Ha. But the man who said it (Colin Chapman) built chassis' that were well known and regarded for working despite his letting them work. He would do that.

Jay Novak, Ford Racing Engineer, writes that with spherical rod ends and near ground roll centers there not enough difference to worry too much about Force Based Moment Centers.

Elsewhere, maybe on Eng-Tips, I read that one of the things that factors into the difference is torsional resistance in bushings. Which doesn't appear to bear any relation to RC height.

I always mention my interest in Jacking Forces. Above ground RC's jack the chassis UP, and below ground RC's jack the chassis down (a friend of mine who's been sceptical about negative jacking forces recently asked Bill Mitchell about them - Bill told him to believe it).

Above ground RC's reduce the springs burden in roll. Below ground RC's increase the springs burden in roll. The RC location doesn't much affect the springs burden in bump.

There's a bit of an orthodoxy that you want to run as soft as you can to maximize mechanical grip. That supports the idea of using the RC location for roll stiffness. But is it worth paying for with jacking forces?

A typical tin top driver can detect a change in the balance of the car from as little as 20lbs of low velocity force from a damper. That's within the scope of available jacking forces from even relatively low above ground roll centers.

And are you really sure you want to run as soft as possible anyway? I think Ren made alot of sense on the DSR board - the spring rate of your tires will dictate your maximum effective wheel/spring rate. And stiffer lets you reduce your static camber compensation, and makes more effective as a percentage your camber change from caster during turn in.

As ever, it all comes down to how happy are the tires...and how much of the time. And how hard you're pushing down on which ones.

It takes about the same amount of raw brain power to solve the problem regardless of approach. A simulation program won't make up for insufficient understanding, and without sufficient understanding a testing program is a waste of time and money.

So am I lost in the woods? If I were to get out of the woods where would I be?

Maybe I'd be in another forest. The "Warp Stiffness" forest?

Fuggin Warp Stifness!

Being lost is one thing. Being lost and having a headache is quite another.

Increased spring rates give increased warp stiffness. Increased Sway Bar stiffness increases warp stiffness. While it IS possible to partially decouple Roll Stiffness from Warp Stiffness, you still have to have a SOFT CAR for it to mean much. And a soft car must necessarily be a high car. And you really can't "decouple" wheel rate from wheel travel.

I keep weighing the comments of the Citroen drivers against my vague conspiracy like theories of who's doing what how. I "think" that as wheel rates go up, and ride heights come down, the potential benefit of this partial decoupling diminishes. But then I ask myself whether or not that's a reasonable idea. Less absolute magnitude of force - ok. Less percentage as compared to other potential improvements - no clue. I can't forget the argument I just made about roll center height.

Scott, who, if forced, will settle for being ignorant and fast...Doesn't sound right does it?

 

help ! slurpy brain freeze


I guess my only question that I can ask is,

In what way are you trying to apply this ?

to a production built car for track racing, or other ?

 

I need to read this in-depth before I say anything. But we're still messing with the suspension and chassis design of our FSAE car so this kind of stuff keeps me awake at night frequently

 

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by RR98ITR &raquo;</TD></TR><TR><TD CLASS="quote">Look at a head on picture of a turning car. The chassis has rolled to an angle - about some axis - presumably a dynamic instantaneous one. At each point of the rotation from rest the locations of the suspension pivots are displaced - giving the instantaneous values of dynamic cambers, cornerweights, and chassis location.</TD></TR></TABLE>

A more detailed model will tell you what the actual vertical load is on the tires, the camber as the body rolls and suspension moves up or down, and the change in toe front and rear for each of the wheels. Then you are going to need a tire model or all your time is wasted, since you cannot calculate the lateral forces, which is all you care about. (You might even need to know the centre of pressure location and lift/downforce coefficients of the car). I have been enjoying (well maybe not enjoying) reading Olley's book recently put out by the Millikens/SAE. In it the Millikens make a note of Olley's work being a little out of date about roll in that force based analysis is a little more current. But a least Olley works with a tire model to look at the Oversteer/Understeer issue as the g's build up in a corner. He even plots the force on each of the four wheels to show how the cornering proceeds. Very nice plots of slip angle front and rear and of them crossing over for a car which starts in understeer at low g's and "converts" to oversteer. Then he proceeds to add the effects of traction which reduces lateral grip of tire. Most interesting, even if not "Force Based Moment Centres". You need a tire model first, or else it is all about guessing.

In the meantime I'm having a lot of fun with my new simulator (written by a guy who works for Pacjeka) that I got for free. If you have an estimate of the heights of the track at Laguna let me know. The supplied tracks are flat but you can edit the height. Have fun with this:

LapSim free edition (it has a nice tire model!):
http://www.bosch-motorsport.co...x.htm

 

i've given up concerning myself with roll centers. i can't change them in my class anyhow. if i did make a change, i wouldn't know in which direction to go to acheive what goal. for now, i'm working with the "leave it the **** alone and drive it" approach. if nothing else, it's cheaper and i spend less time under the car.

nate

 

Claude,

I'm just a caveman. My primitive mind is not capable of comprehending a "simulation" that is "possessed" by spirits (slip angles, lateral force vectors at the contact patches, and such).

I can move chassis end pivots up to 25mm.

I haven't the slightest interest in a numerical approach to the problem. I have a considerable interest in a spatial/functional approach to the problem. Think about how you take lap times or tire temps from the hot pits. You're looking at functional indicators, and thinking in terms of "getting better" or "getting worse".

It may sound naive and stuperficial for me to express my desire to keep things primitive, but I'd argue the opposite - I'm trying to work successfully within my limitations by refusing to get entangled in complexity.

I've thought about ordering Olley's collection, but haven't wanted to spend the money on something I feared would resemble Bastow's work - meaning highly numeric. Is there much conversational/topical content?

Scott, who saw a big book on Birel the other day...looks like a good fun read...

 

Scott, who, if forced, will settle for being ignorant and fast...

This Scott doesn't need to be forced.

 

Geometry is only the starting point. whether done in 2D or 3D, you get the roll center position and then you can calculate the weight transfer. You can also get the camber/toe curve. You can see how moving a suspension point affects this, in terms of keeping the toe/roll effects to what you think is good. How much roll understeer do you want at the rear vs.the front. Tire temps won't tell you anything about this. Adjusting tire temps (the stuff you can measure in the hot pits) will only tell you if you have optimized the current setup. Aside from static camber changes to be made, it won't tell you much else about what to do to the car, but it tells you if you are going in the right direction. If you do camber sweeps on a new tire keeping all else constant, then you can learn about what camber your tire likes best.Change the front and then change the rear. Then look at tire temp effects. But you will still not know what to do with the roll centre.I think tire temps is a very good way to see if something is off, too much camber, too much roll or too much weight transfer at one end, but you can have tire temp profiles looking good and not know what to change to make it faster. I think this is where analysis and simulation comes in. Once you are pretty fast and want to go faster, it is getting harder and more work is going to be required. For example, I don't imagine anyone is running at the sharp end of the World Challenge without data logging. They've had it in the RTR cars for 8 or more years now. While force based moment centre calculations may not be accessible for now, data logging is. I have turned on a few friends of mine to it, and now they never leave home without it. Some just observe, while others analyze. GPS system gives you sector times for any sector you care to define. You will know where you are fast and where you are slow when you make a change. Tire temps won't tell you that.

With regards to Olley's book, it is definitely pretty mathematical,but it fills in the blanks from the Milliken book, where you have sometimes no idea where they got stuff. Looks like they used Olley's work as the basis for their book. It has given me much better appreciation of the workings of slip angles, and lateral forces arising from them. Practically the only book that treated Understeer/Oversteer based on a tire load curve was Herb Adam's book. All this geometry stuff without tire load curves is not of much use. You needto know the efet on "How much grip have I got now?". Now the Olley book has just improved things by making slip angles more friendly than before. In particular I would recommend chapter 2 in the book.