Geometry Lessons

Bump and Roll

I've been curious for a long time about bump steer and just how much has been built into these old cars.  Bump steer if you didn't know, is the change in front tire toe-in or toe-out as a function of the up and down (bounce and rebound) front A-arm angles.  In a straight line on a bumpy road, bump steer will cause the car to randomly wander.  Generally the designed in geometry increases toe-out as the A-arms lift in relation to the frame and this in turn increases understeer.  The outside wheel in a turn moves up in relation to the frame and being toed-out, doesn't turn as tight.  It "plows" or "pushes" a little.

Closely related to bump steer is roll steer which is toe-change as a function of vehicle roll angle.  If toe-out on one side doesn't equal toe-in on the other side there is a steering effect due to vehicle roll.  This is important in autocross because of the tight turns at speed.

Too much toe-out in whatever situation will cause a tire to scrub and that lessens its grip and increases drag.  In the simplest terms, the car is having to push the tire around a turn rather than having it roll in the intended direction.  Bump and roll are undesireable attributes in autocrossing and it is in my best interest to remove these effects to the extent that they are significant.  While I can find plenty of discussion on the internet on these topics I haven't found any measurements that give me a seat of the pants feel for the magnitude of change that's been designed in.

As of today, all of my front suspension components sans coil springs have been bolted in place and snugged up (but not yet torqued).  I've finally gotten completely away from the binding associated with rubber and poly bushings and the A-arms move up and down smoothly with minimal effort.  Further there is likely to be almost no compliance as a function of braking or accelerating.  Compliance in this case would be the A-arms twisting about in the bushings also unpredictably affecting toe.  This too unsettles the driver (me) because the car doesn't seem to be going exactly where I want it to.  Modern sports cars (e.g. the Z06) feel so much more precise and this is what I'm after in this rebuild.

On hand today are my el-cheapo chinese lasers, a stock tie-rod setup and a bump-steer kit from vette brakes and products.  I've set the toe to zero at full bounce using a machinist's square held against the spindle's axle shaft and the long iron bar I used previously laid along the frame.  Close enough for now.  This is a crude but known starting point for relative order of magnitude measurements.  Remember the pareto principle (also known as the 80-20 rule) states that, for many events, roughly 80% of the effects come from 20% of the causes.  So I figure I am about 80% right having done about 20% of the work.  The big money boys that charge for these alignments start with the exact dialed in caster camber and toe and everything bolted up with the car assembled and sitting on its wheels.  And some fancy bump steer gauges.

Here to put pictures to these words is what the A-arm looks like.

This photo is taken under the frame looking up at the left lower control arm.  The arm swivels about the axis defined by the aluminum crossmember labelled GLOBAL WEST which is bolted to the  frame.  The tie rod is to the left of the control arm.

Close up of one of the control arm delrin bushings.

The setup is once again to stick a laser onto the car and plot where it points.  Data collection is trivial at this point.  Lift up on the axle with one hand (no coil spring, remember and the bushings move freely) and reach over with the other hand and mark with a felt pen where the laser is pointing.  The box is 32" from the ball joint.  This gives me a little more than twice the actual toe, given that the tire radius is about 15".  Mark a few points until I've gone through the full range of travel and then do the other side.  The steering input does not move during this process so the steering angle is solely dependent on what the tie rod position is relative to the steering arm and control arm.

We are now on the left side of the car looking forward.  The laser is fastened with plumbers putty and you can just see the red light shining onto a large cardboard box near the crease on the left hand side.  These cardboard boxes seem to arrive daily.  This one had my SS brake and gas lines.  It is wider than the car.  Similar situation on the right hand side.

 Here are the results for no bump steer compensation plotted in black and the bump steer kit installed and plotted in red.

Left A-arm. 

Right A-arm

In motion on a track the wheels would never move over this range.  This is essentially rubber bumper to rubber bumper.  They would however move a few inches somewhere near the center of this range.  Since I don't have a complete car on the ground I don't really know where the middle is.  Even so, the slope of black uncompensated curve is about 1 inch of toe for 3 inches of up and down travel.  Half that distance would be a toe change of 1/2 inch per wheel.  A typical autocross specification for toe would be about 1/8 inch so this is a rather large unwanted steering input.  All in all I'm quite surprised in how bad the toe is in stock form. 

With the bump steer kit installed things look pretty good.  The RHS is almost perfect.  The LHS is pretty good.  What may be affecting that side is that I have not torqued down the Pittman arm on the steering box and it is sitting about 3/8 inch lower than what should be its final position.

The kit is pretty simple, a steel block bolted to the steering arm and shorter tie rod sleeves.  The block lowers the ball joint and moves it inward

Back View

Top View

This is kind of neat to play with by hand.  The laser is stuck near the lower ball joint pointing straight ahead as it is swept through the range of the arm.  Naively you would expect that the trace should describe a small arc on the cardboard box because that is what the control arm is doing.  That's not what really happens, as the up and down straight line trace proves.  It's doing what it is supposed to.  As the tie rod moves with the control arm it steers the spindle as well and the resultant of both motions is zero change in toe.  Truth be told I haven't completely visualized this.  But to complete the picture here is the set of constraints which the tie rod must meet.

Theory

Compare to the real thing!

The steering box interferes with the view on the LHS so this is the RHS.  The pivot for the upper control arm can just be seen over the fuel lines.  The upper ball joint is barely visible as well, in the gap between the fuel lines and the frame.  Also perspective intrudes here a bit.  Things don't quite line up with theory given Chevy's design decisions, but they are evidentally close enough to give excellent results.

Ackerman Steering

Drawing taken from Wiki Commons

Now for the "unintended consequences"!  Any change usually has some.  The bump steer kit modifies the steering geometry, specifically affecting something called Ackerman steering.  The front wheels of a car making a turn are on different circles.  The inside wheel is on a tighter radius than the outside wheel.  Ignoring effects such as tire slip angle, the wheels should roll freely along the inner and outer circles.

Approximation for Ackermann geometry

The geometry of the steering arms with Ackerman steering is nominally arranged so that the front wheels axes intersect on the axis of the rear wheels.  This can be determined pretty well with the steering set straight ahead.

Here is the top view of the left hand steering arm with the wheels pointed straight ahead.

The string tied to the lower ball joint and running across the castle nut on the bump steer kit is used to determine the actual location of the turn circle.  The string here intersects a similar string on the other side of the car.  This is where the the center of the turning circle is in relation to the four wheels.  It may or may not be at the rear axle.

The stock tie rod ball joint position in the steering arm is the left hex bolt.  The tie rod, castle nut, and old position are in a straight line. For small steering corrections we would expect no difference in handling. 

For tighter turns, things have changed.  The tie rod no longer forms a straight line with the stock ball joint position.  There is some loss in steering ratio for the outside tire.  It will seem to the driver that he has to turn the steering wheel more for tighter turns.   The inside tire meanwhile is trying to travel on a tighter circle. These pictures may help and it may also help to imagine the projection of the tie rod in relation to the stock position (left bolt).

Left wheel position in a right hand turn
This is the heavily loaded tire

Left wheel position in a left hand turn
Lightly loaded with increased toe out

This quote from the internet gives me some feeling for how to think about this aspect of steering geometry in autocross

• the reason for using ackerman in racing is different, and specifically it is different in autocross than in road course or other types of faster racing. this is because we use much larger steer angles than any road course car.
• because all the tires of a race car are operating at some slip angle in a corner, the specifics of ackerman geometry are less important than for parking maneuvers.
• what matters for us is that positive ackerman steering means the front toe out increases with steer angle. for less than large steer angles the amount of toe change is very small.
• so the question becomes: what effect does increased toe-out have on cornering?
• in a corner, there is a more heavily loaded outside tire and more lightly loaded inside tire. the outside tire also (should be) optimized in camber and will (should) be operating at optimal slip angle to generate maximum lateral force.
• this means that the inside tire is more lightly loaded, and operating at a higher slip angle than the outer tire.
• the key point is: this high slip angle inner tire creates significant drag.
• this drag creates a turning force, much like the stability system works on a modern car by applying the brake on just one side of the car and keeping you on the line the computer detects that you want from reading the steer angle.
• since the increased drag robs acceleration, positive ackerman may work best on over powered cars like the cobras.

I can't really take an overhead photo but this is the projection

Seems right!  Given all this theory I was interested in finding out what the Stingray designers did, and what I changed.  So I ran two strings as shown in the picture to the back of the car to see what the bump steer modifications did to the stock design, fully expecting that the kit had fixed one thing at the expense of another.  Not so.  Despite the fact that Ackerman steering was patented in 1817 for horseless carriages, there is little to no Ackerman compensation in the design (assuming I am doing all this right).  I was kind of surprised.  The stock Corvette suspension for my car and the C2/C3 series was taken from the Chevy production parts bin to save costs.    What money was available was used to move from a solid axle to an IRS, for which I am grateful.

The good news is that the bump steer moves the turning circle from way behind the car to somewhere under the gas tank.  While this is not quite at the rear wheels, if one considers the effect of the slip angles, it turns out that turning circle moves forward still more and I will have the outside tire grabbing more traction, as in the quote.

Slip angle is the additional angle req'd to exert a turning force.  The tires may be pointing in one direction but they are rolling in a slightly different direction.

And there's always torque induced oversteer (which we Corvette drivers love) to help get me around turns.  It's probably why these cars do as well as they do despite their primitive technology.  It's the American way, more is better!

I learned a bit in this exercise.  I guess I could have just bolted things on and hoped the handling improved but at least I have a better idea of what to expect and what to look at.  This took a bit of time but I'm still waiting on my differential.

Next up the rear supension.  New parts there too!