Bump Steer at Eldora vs Your Home Track

Bump steer is the single most misunderstood geometry problem in dirt racing, and the reason is simple: every article you have ever read about it was written for pavement. Wikipedia gives you 120 characters and a diagram of a tie rod arc. That is cute. It tells you nothing about what happens when your left front drops 3 inches into a hole off turn two at Eldora while the banking is pushing 24 degrees of lateral load through your spindle at 140 mph. Bump steer on dirt is not a design flaw you eliminate. It is a variable you manage — and on heavy clay, with ruts, with banking transitions, with moisture changes lap to lap, it becomes the single biggest source of the "my car suddenly started pushing" complaint that fills my pit area every Saturday night.

What It Actually Is — The 30-Second Version

Bump steer is toe change caused by suspension travel. That is it. When your wheel moves up (compression) or down (droop), the tie rod and the lower control arm trace different arcs because they have different lengths and different pivot points. If those arcs do not match, the wheel toes in or toes out as it travels. The wheel steers itself without the driver turning the wheel. On pavement, where suspension travel might be 1.5 inches in a corner, the effect is small and predictable. On dirt, where the left front alone can see 3–4 inches of travel through a single corner at Eldora — and 1.5–2 inches at your flat local 3/8-mile — bump steer is not a subtle geometry curiosity. It is a handling condition that changes every time the track surface changes.

The measurement is simple: inches of toe change per inch of vertical wheel travel. A common target on pavement is 0.030 inches of toe change per inch of travel. On dirt, where travel is doubled or tripled, that same 0.030 per inch becomes 0.090–0.120 total toe change through the stroke. That is enough to make the car wander, push, or snap loose depending on direction.

Why Eldora Is the Graduate-Level Exam

Eldora Speedway is a half-mile, banked 24 degrees in the corners, and the surface is famous for developing deep cushions — 4 to 6 inches of loose material stacked against the wall — and ruts that carve 2–3 inches into the racing surface by the B-main. The track crew works it between races, but by the feature the groove is a violent place. Here is what that means for bump steer specifically:

A 410 sprint car at Eldora with standard ride heights — LF at 4.5–5.5 inches static — can see the left front compress 3–4 inches through the corner from combined banking load, wing downforce (400–800 lbs depending on angle), and rut displacement. At the same time, the right front may droop 1–2 inches as weight transfers left. That is 5–6 inches of total differential travel between your two front wheels. If your bump steer curve is not nearly flat through that range, each wheel is toeing a different amount in a different direction. Your effective toe setting mid-corner has nothing to do with what you set in the pits with a tape measure and a straight edge.

Compare that to your home 3/8-mile flat track — maybe 6–8 degrees of banking, smoother surface, lower speeds, shorter features that do not chew the clay as deep. Your LF might see 1.5–2.5 inches of compression. Your RF droops an inch. Total differential travel is 2.5–3.5 inches. Same car, same geometry, completely different bump steer expression. The setup that felt neutral at home pushes at Eldora. Not because the springs are wrong. Not because the torsion bars are wrong. Because the tie rod arc does not match the A-arm arc through twice the travel range.

The Geometry That Causes It

The tie rod connects the steering rack (or drag link) to the spindle. The lower control arm connects the chassis to the spindle. Both pivot. Both trace arcs. The tie rod arc has a radius equal to the tie rod length from its inner pivot to the outer ball joint. The lower arm arc has a different radius — the arm length from its chassis mount to the ball joint. If these two arcs diverge as the suspension moves, the spindle rotates around its kingpin axis. That rotation is toe change. That is bump steer.

Three variables control the curve:

1. Tie rod height at the spindle. Moving the outer tie rod end up or down changes where the tie rod arc intersects the arm arc through travel. This is the most common adjustment. Shims under the tie rod end — 0.060-inch shims are standard, and most teams carry a stack from 0.030 to 0.250 — raise or lower the outer pivot to reshape the curve. One 0.060-inch shim typically changes bump steer by 0.008–0.015 inches per inch of travel. That sounds small until you multiply by 4 inches of travel at Eldora: 0.032–0.060 inches of toe. Enough to feel. Enough to matter.

2. Tie rod length. Longer tie rod = larger arc radius = flatter arc through the same travel. This is why adjustable tie rods exist. But length also affects Ackermann geometry (how much the inside wheel toes out more than the outside wheel toes in during steering input), so you cannot just make the tie rod infinitely long without screwing up your turn-in behavior.

3. Inner pivot height. On a rack-and-pinion car, this is the rack height. On a drag-link car (most sprint cars), it is the drag link mount on the steering arm. Moving the rack up or down changes the tie rod angle, which changes the arc. Rack height is the coarsest adjustment — 1/4 inch of rack height change can move the bump steer curve dramatically, 0.020–0.040 inches per inch of travel. Most sprint car chassis have slotted rack mounts that allow 0.5–1.0 inch of vertical adjustment.

Class-by-Class: Where the Physics Diverges

This is where every pavement-derived article falls apart. Different dirt classes have fundamentally different front suspension architectures, and bump steer manifests differently in each one.

410/360/305 Sprint Cars — Torsion Bar Front

Sprint cars run torsion bar front suspension with unequal-length A-arms. The LF torsion bar is typically 850–1050 lb/in, the RF 925–1200 lb/in. Caster split is significant: RF 4–7 degrees, LF 0–2 degrees. That caster split means the two sides have different kingpin axis orientations, which means bump steer affects each side differently even with identical tie rod geometry. The right front with 6 degrees of caster has more kingpin inclination effect on toe than the left front with 1 degree. Most teams set bump steer with the same shim stack on both sides. This is wrong. You need to check and set each side independently.

Sprint cars also have no front brakes. Zero. The only brake is LR. This means the front suspension is never loaded by braking forces — only by banking load, wing downforce, and surface impacts. The bump steer curve only matters in compression under cornering load (left side) and droop under cornering unload (right side). You do not need the curve to be flat through the full travel range. You need it flat through the loaded range on each respective side.

Common mistake: teams set bump steer at static ride height and check 1 inch in each direction. At a flat home track, that covers most of the working range. At Eldora, you need to check 3.5 inches of compression on the LF and 2 inches of droop on the RF. The curve might be dead flat from 0 to 2 inches and then toe out 0.040 inches from 2 to 3.5 inches. You would never see that problem at home.

Late Models — Coil Spring Front, and the Fifth-Coil Complication

Late models run coil spring front suspension with upper and lower A-arms. RF spring rates run 700–900 lb/in on supers, 600–800 on crates. LF runs 650–800 and 550–700 respectively. These are softer than sprint car torsion bars, which means more travel for the same load. A super late model at Eldora can see 3–4 inches of LF compression — similar to a sprint car — but the car weighs 2,300 lbs versus 1,400, so the energy stored in that compression is much higher and the rebound behavior is different.

Late models also have front brakes. This loads the front suspension under deceleration into the corner, adding a travel phase that sprint cars do not have. Your bump steer curve needs to behave during braking compression AND cornering compression. Those are sequential events with the load shifting from bilateral (both fronts loaded equally under braking) to asymmetric (LF loaded, RF unloaded in the corner). The bump steer that felt neutral under braking might toe-out during cornering because you are now 1.5 inches deeper into the travel on the LF only.

The post-2015 revolution — the fifth coil and torque arm rear suspension — introduced rear steer as a tuning tool. Rear steer can mask or compound bump steer effects. If your rear is steering the car left under power (tightening) and your front bump steer is toeing out under compression (loosening), they might cancel out and the car feels neutral. Change one without understanding the other and the balance falls apart. I have seen teams chase a push for three races by adding front spring rate when the actual problem was 0.060 inches of bump steer toe-out at 3 inches of compression. They stiffened the spring, reduced the travel, and reduced the bump steer expression without ever finding the cause. Then they went to a rougher track and it came back worse because the travel exceeded the spring's ability to control it.

Modifieds — The Harris Torque Link Variable

IMCA-style modifieds on GRT or Harris frames weigh 2,400+ lbs with spring rates of 800–1,200 lb/in. The front suspension is conventional A-arm/coil spring, and bump steer setup is similar to late models. But the defining chassis feature — the Harris torque link rear — creates a rear suspension behavior that interacts with front bump steer in a unique way. The torque link controls rear roll stiffness and rear steer simultaneously. If the rear is steering the car and the front has bump steer, the two inputs can fight each other mid-corner, creating a condition drivers describe as "the car changes direction halfway through the turn." That is not the track. That is geometry disagreement between the front and rear.

Street Stocks — The Chrysler Ball Joint Secret

Street stocks on GM metric chassis are running stock-type front suspension. The classic upgrade — Chrysler ball joints on GM spindles — changes the geometry dramatically. Chrysler upper ball joints sit at a different angle than GM, which changes the steering axis inclination and the kingpin angle. This moves the tie rod arc relative to the arm arc. Teams that bolt on Chrysler joints without rechecking bump steer are introducing an unknown quantity of toe change through travel. I have measured street stocks with 0.080 inches of toe-out per inch of compression after a ball joint swap. On a stock GM spindle with stock joints, the same car showed 0.025. The Chrysler joints are the right move for front-end durability and caster adjustability — but you must reshim the tie rods after the swap.

Micro Sprints — Wishbone vs. Beam Front

Wishbone front micros (Hyper X6, premium chassis) have independent A-arms and full bump steer adjustability through shims and tie rod length. Beam axle fronts (Stallard SST, older Kiwi, Z-link) have a single tie rod connecting the steering to the beam, and bump steer is a function of the beam pivot angle relative to the tie rod mount. On beam axles, you have less adjustment range — you can shim the tie rod end and adjust the tie rod length, but you cannot independently set each side. Both wheels are connected, so a bump on one side creates a bump steer input on both.

Micro sprints weigh 800–1,000 lbs with typical front ride heights of 3–4 inches. At 6–10 psi tire pressure, the tire itself absorbs some impact before the suspension moves. Total LF suspension travel on a micro is typically 1.5–2.5 inches, which puts bump steer sensitivity in the moderate range. But the car is so light that small toe changes have proportionally larger effects on direction — 0.030 inches of bump steer toe-out on a 900 lb micro has more directional effect than 0.030 on a 2,300 lb late model. Light cars amplify geometry errors.

Intentional Bump Steer — The Trick Nobody Writes About

Here is where the pavement knowledge completely fails you. On asphalt, the goal is zero bump steer. Flat curve. No toe change through travel. The surface is predictable, the loads are predictable, and any unwanted steering input is pure noise.

On dirt, the surface is not predictable. And the fast guys use that.

A small amount of toe-in on left-front compression — 0.005 to 0.010 inches per inch of travel — tightens the car when the LF loads deepest. That happens at the bottom of ruts, at the apex of high-banked corners, at the moment of peak lateral load. The car self-tightens at the exact moment it would otherwise push from excess front-end load. This is not accidental. This is deliberate geometry tuning for dirt conditions.

At Eldora specifically, where the LF can compress 3.5–4.0 inches through the big end, teams running 0.008"/in toe-in on compression get 0.028–0.032 inches of total toe-in at peak travel. That is barely perceptible in a static measurement but absolutely felt at 140 mph through a rut. The car rotates slightly at the apex without the driver adding steering input. The driver describes it as "the car turns itself." It is not magic. It is geometry doing exactly what someone set it to do.

Conversely, at your flat home track where the LF only compresses 1.5–2.0 inches, that same 0.008"/in toe-in produces only 0.012–0.016 inches of total effect. Barely noticeable. The car that "turns itself" at Eldora feels normal at home. This is why setup notes from big races rarely translate directly to local tracks — the bump steer effect scales with travel, and travel scales with banking, surface roughness, and speed.

How to Measure It — The Shop Procedure

You need: a bump steer gauge (Longacre makes a solid one for $180–$250), a set of bump steer shims (0.030, 0.060, 0.090, 0.125 — most kits come with an assortment), a jack, two jack stands, and 45 minutes per side. That is it.

Step 1: Set the car at ride height on jack stands under the frame. Remove the spring or disconnect the torsion bar so the spindle hangs free. The steering must be locked — tie the wheel or clamp the column.

Step 2: Attach the bump steer gauge to the rotor or hub face. The gauge has a flat plate that bolts to the rotor studs and a dial indicator that reads toe change as the spindle moves vertically.

Step 3: Use a floor jack under the spindle to raise the wheel to full compression (wheel up into the fender — whatever your bump stop allows). Zero the gauge. Then lower the jack slowly, reading the dial indicator every half-inch of travel. Record: ride height, 0.5" compression, 1.0", 1.5", 2.0", 2.5", 3.0", 3.5". Then continue into droop: 0.5" droop, 1.0", 1.5", 2.0".

Step 4: Plot the numbers. You are looking at a curve. A flat line means zero bump steer. A curve that bows toward toe-in means the wheel toes in during that phase of travel. Toe-out, the opposite. Most real-world curves have a shape — they might be flat from 0 to 2 inches and then curve toward toe-out beyond that.

Step 5: Adjust with shims under the outer tie rod end. Adding shims (raising the tie rod end) typically moves the curve toward toe-out in compression. Removing shims moves it toward toe-in. One 0.060" shim moves the curve 0.008–0.015"/in — but this varies by chassis, spindle, and tie rod length. You must re-measure after every change.

Common mistake number one: measuring only at ride height plus and minus 1 inch. That is a pavement measurement. On dirt, you need the full range. Common mistake number two: measuring with the spring or torsion bar connected. The spring adds friction and hysteresis that mask the true geometry. Disconnect it. Common mistake number three: not locking the steering. Any steering column movement during measurement invalidates the reading.

The Track Surface Variable That Nobody Accounts

SCALE SHEET WITH LEARNING

Enter your four corners. Black Book calculates cross, left, rear, bite — color-coded for your class. Log adjustments and it learns how YOUR car moves weight.

TRY THE SCALE SHEET FREE