A ball differential is a torque-biasing differential that uses a ring of small hardened balls trapped between two flat drive plates, clamped by an adjustable spring, to allow left and right output shafts to rotate at different speeds while transmitting drive torque. Associated, Yokomo, and Tamiya 1/10 scale touring and pan cars use it as the primary rear diff. The clamping spring sets the slip threshold, which lets the driver tune corner-exit traction by hand. The result is a sub-30 gram differential with adjustable torque transfer that gear diffs cannot match for tunability.
Ball Differential Interactive Calculator
Vary clamp preload, friction, ball track radius, ball count, and drive torque to see slip torque, load sharing, and slip margin.
Equation Used
This simplified ball differential calculator treats the diff as two friction interfaces acting at the ball track radius. Increasing clamp force, friction coefficient, or track radius raises the static torque before slip. A positive torque margin means the applied drive torque is below the calculated slip threshold.
- Simplified dry-friction torque capacity model.
- Clamp force is total axial preload through the diff stack.
- Two effective friction interfaces act at the same ball track radius.
- Friction coefficient is constant; wear, heat, and surface damage are not modeled.
Inside the Ball Differential
The mechanism is simple in principle and unforgiving in execution. You have two parallel diff plates — usually hardened tool steel or tungsten carbide — facing each other on a common axle. Between them sits a cage of typically 10 to 14 balls (3/32 inch carbide is the standard in 1/10 scale) running on a circular track machined into each plate. One plate is keyed to the spur gear or pulley, the other plate drives one output shaft directly, and the opposite output shaft is keyed to the through-axle. A Belleville-washer stack or coil spring squeezes the whole sandwich together through a thrust bearing, and that clamp force is what sets the static torque the diff can transmit before it slips.
When both wheels see equal load, the balls roll around with the plates as a solid unit and the diff acts like a spool. The instant one wheel meets more resistance — say the inside rear loading up under throttle on corner exit — the higher-traction side pulls harder on its plate, the balls start to skid microscopically across the plate face, and torque biases toward the slower-turning wheel. That skid is where ball differentials live or die. If the plate face is pitted, scored, or has lost its mirror finish (the surface should be Ra 0.2 µm or better), the balls dig in, the action becomes notchy, and the diff either grabs on-throttle or refuses to hold drag brake. If the clamping spring backs off — which happens because the plates wear flat — slip torque collapses and you'll see the diff smoke out of slow corners.
Build tolerances are tight. The balls must be matched in diameter to within 2 µm or you get preferential loading on the high spots. The thrust bearing has to spin freely under preload — a sticky thrust race feels exactly like a worn diff and fools a lot of builders. And the diff plates must be re-faced or replaced once you can feel a ridge with a fingernail; carbide balls are harder than the plates, and they will groove the surface long before the balls themselves wear out.
Key Components
- Diff Plates (2): Hardened steel or tungsten carbide discs, typically 13-16 mm OD, that the balls roll between. Surface finish must be Ra 0.2 µm or better, and flatness within 5 µm across the face. One plate keys to the drive pulley, the other to one output shaft.
- Diff Balls: Tungsten carbide balls, 3/32 inch (2.381 mm) is the 1/10 scale standard, in a 10-14 ball ring. Hardness is typically 90-92 HRA. Diameter must be matched within 2 µm across the set or load distributes unevenly.
- Thrust Bearing: Small thrust race (usually 5x10x4 mm) sitting between the clamp spring and the rotating diff body. Carries the full preload — about 30-80 N depending on setup — while letting the spring stay stationary relative to the chassis.
- Belleville Washer or Coil Spring: Provides the clamping force that sets slip torque. A typical 1/10 ball diff runs 1.5-3 N·m of slip torque, and turning the adjuster nut 1/8 turn changes that by roughly 0.2 N·m.
- Adjuster Nut and Drive Hub: Threaded female component the user turns to compress the spring. The drive hub on the opposite end keys to the output shaft and also locates the through-axle.
- Through-Axle: Hardened shaft, typically 5 mm with a flat or pin, that passes through the centre of the diff and drives the opposite wheel. Concentricity to the diff plates within 0.02 mm or you get vibration at speed.
Industries That Rely on the Ball Differential
Ball differentials show up wherever a designer wants adjustable torque biasing without the weight or cost of a planetary gear diff. The dominant home is small-scale RC racing, but the principle has shown up in scientific instruments, high-end model engineering, and a handful of small mobile robots where slip behaviour matters more than absolute torque capacity.
- RC Racing — Touring Car: Team Associated TC7.2 and Yokomo BD11 1/10 electric touring cars use ball diffs front and rear, with carbide balls and tungsten plates as the standard race-spec build.
- RC Racing — Pan Car: Associated RC10F6 and CRC Gen-XL 1/12 pan cars rely on a single rear ball diff as the only differential, since front wheels are independent. Slip torque is tuned per track.
- Off-Road RC: Tamiya TT-02 and TA08 kits ship with ball diffs in entry-level form, though most pro 1/8 buggies have moved to gear diffs for shock loading reasons.
- Educational Robotics: FIRST Robotics and university coursework drivetrains occasionally use ball diff cartridges for compact differential drive on demonstrator platforms, particularly where torque biasing teaches a clear lesson.
- Precision Instruments: Older mechanical analog computers and some film-camera autofocus drives used miniature ball-and-disc differentials for the same torque-biasing reason — Bowmar and early Friden machines used variants of the principle.
- Small Mobile Robots: Sub-2 kg research robots from groups like MIT Leg Lab have used ball diff cartridges where weight is critical and shock load is low.
The Formula Behind the Ball Differential
The number practitioners care about is slip torque — the torque the diff transmits before the balls start skidding across the plates. It comes from the clamping force, the ball ring radius, and the friction coefficient between ball and plate. At the low end of typical clamp force the diff slips on hard acceleration and drags the inside wheel like a wet clutch. At the high end it locks solid and behaves like a spool, killing corner entry. The sweet spot for a 1/10 touring car sits around 2 N·m, just tight enough that you can pinch the right rear wheel between two fingers while the left rear rotates the spur gear without grunting.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| Tslip | Torque required to make the balls skid on the plate face | N·m | lb·in |
| μ | Effective friction coefficient between carbide ball and plate face (typically 0.10-0.15 dry, 0.04-0.07 with diff lube) | dimensionless | dimensionless |
| Fclamp | Axial clamping force from the Belleville stack or spring per ball contact | N | lbf |
| rring | Radius from diff axis to centre of the ball ring | m | in |
| ncontacts | Number of effective ball-plate contacts (number of balls, since each ball contacts both plates symmetrically) | dimensionless | dimensionless |
Worked Example: Ball Differential in a 1/10 scale carpet touring car ball diff
You are setting the rear ball diff on a Yokomo BD11 1/10 touring car for a high-grip carpet round at a regional ROAR event. The diff has 12 carbide balls of 3/32 inch on a 10.5 mm ring radius, runs Associated Stealth diff lube, and the spring stack delivers an adjustable axial preload from 20 N to 80 N total. You want to know what slip torque you'll see across the adjuster range and where to start.
Given
- nballs = 12 balls
- rring = 0.0105 m
- μ = 0.05 (lubricated carbide-on-tungsten)
- Fclamp,total = 20 to 80 N
Solution
Step 1 — at the nominal mid-range preload of 50 N, the per-ball clamp force is the total preload divided by the ball count:
Step 2 — compute nominal slip torque using the friction-radius formula. Each ball contributes one effective slipping contact (the plates skid relative to each other through the ball):
That feels light because the formula is computing torque at the ball ring radius. Multiply by the spur-to-axle ratio (typically 1:1 internal in a ball diff) and you've got the slip torque seen at the wheel — about 0.32 N·m, which is the right ballpark for a 1/10 carpet diff that holds drag brake but releases on corner exit.
Step 3 — at the low end, 20 N preload:
That is too loose. The diff will smoke out of slow hairpins and you'll feel the inside rear spin up the moment you crack the throttle. Drag brake is essentially gone — coast into a corner and the car pushes wide because the diff freewheels.
Step 4 — at the high end, 80 N preload:
Now the diff behaves nearly like a spool. The car will rotate cleanly on power but refuse to turn on entry — you'll see understeer at turn-in and the rear tyres will scrub audibly through long sweepers. The sweet spot for 27.5 mm carpet rubber is right at the nominal 50 N setting, with the field typically running ±10 N either side based on grip.
Result
Nominal slip torque at 50 N preload comes out to roughly 0. 32 N·m at the wheel — the diff holds light drag brake into the corner, then breaks loose progressively as you feed throttle. The 20 N setting drops slip to about a third of nominal and the diff acts open, while the 80 N setting nearly triples it and locks the rear axle into spool-like behaviour; the 40-60 N window is where most carpet touring cars find their fastest lap times. If you measure or feel slip torque well below predicted, check three things first: (1) thrust bearing balls flattened or pitted, which absorbs preload as drag instead of clamping, (2) diff lube contaminated with tyre additive or motor dust, dropping μ below 0.03, or (3) the female adjuster nut backing off because the threadlock washed out — a classic symptom is the diff feeling perfect for two runs then going soft mid-heat.
Choosing the Ball Differential: Pros and Cons
Ball diffs are not the only way to split torque between two driven wheels, and each alternative wins on a different axis. Here is how a ball diff stacks up against a planetary gear diff (the silicone-oil-filled type used in 1/8 buggies and 1/10 4WD) and a solid spool (zero differential action).
| Property | Ball Differential | Gear (Planetary) Differential | Spool |
|---|---|---|---|
| Slip torque adjustability | Continuously variable via spring preload, 0.1-1 N·m typical range | Set by silicone oil viscosity (1k-1M cSt), changing requires disassembly | Not adjustable — always 100% locked |
| Mass (1/10 scale) | 18-30 g typical | 35-55 g typical | 12-18 g typical |
| Shock load capacity | Low — heavy curb impacts brinell the plates | High — gears spread load across multiple teeth | Highest — single solid shaft |
| Service interval | Re-grease every 3-5 race heats; replace plates every 30-50 heats | Oil change every 10-20 heats; gears last hundreds of heats | Effectively zero — bearing service only |
| Cost (typical 1/10 race-spec) | $40-70 USD with carbide kit | $60-100 USD plus oil sets | $15-25 USD |
| Best application fit | 1/10 carpet/asphalt touring, 1/12 pan cars, smooth-surface racing | 1/8 buggy, 1/10 4WD, off-road with jumps and curbs | Drag racing, oval, drift, traction-limited classes |
| Tuning feedback granularity | Very fine — 1/8 turn of nut shifts car balance noticeably | Coarse — oil viscosity steps are 5k, 7k, 10k cSt | None |
Frequently Asked Questions About Ball Differential
Bench testing loads the diff at low speed and low temperature. On track the plates and lube heat up, the lube thins, and any micro-debris that has worked its way between balls and plates gets dragged into the contact patch. The chatter is stick-slip — the balls grip, release, grip, release at audible frequency.
Pull the diff and look at the plate face under a strong light. If you see concentric scoring rings or a frosted band where the balls run, the plates are done. Carbide balls grooving steel plates is the most common cause; swap to tungsten carbide plates if you're running carbide balls and the issue keeps coming back.
Two questions decide it. First — does the surface have jumps, curbs, or sharp transitions? If yes, gear diff. The shock load through a hard landing brinells ball-diff plates in a single hit and you'll never get the diff smooth again. Second — how much in-session tuning do you want? Ball diffs let you turn a nut between heats; gear diffs need disassembly and oil change.
Carpet touring, 1/12 pan, and indoor asphalt all favour ball diffs. 1/8 buggy, 1/10 short course, and any rough outdoor surface favour gear diffs. The middle ground — 1/10 4WD on smooth astro — splits the field roughly 50/50.
Past the sweet spot, more preload doesn't make the car rotate better, it just removes corner-entry differential action. The car loses turn-in because both rear wheels want to rotate at the same speed mid-corner, the front tyres are dragged sideways, and you scrub speed. Lap time drops by 0.1-0.3 seconds per lap on a typical 1/10 carpet track.
Back off to your previous setting and instead check ride height, rear toe, and tyre additive coverage. A tight diff is often a band-aid for a different chassis problem.
For 17.5T or 21.5T stock racing on carpet, target about 0.25-0.35 N·m at the wheel. The classic field test: hold the spur gear stationary with one hand, grip one rear wheel firmly, and try to rotate the opposite rear wheel by hand. It should rotate with steady pressure — roughly the force needed to crush a soft-drink can — without jerking or popping.
If it rotates freely with light pressure the diff is too loose. If you can't rotate it with a firm two-finger grip it's too tight. Modified classes (10.5T and below) typically run 10-20% tighter because peak torque is higher.
Heat-induced viscosity drop in the diff lube combined with thermal expansion of the spring stack. As the diff warms from 25°C to 60°C through a heat, the Belleville stack relaxes by a few percent and the lube film thins, so effective μ falls. The combined effect can drop slip torque by 30-40% over five minutes.
Two fixes. Use a higher-temperature diff grease (Associated Stealth or Yokomo BB-Grease are formulated for this) instead of plain bearing grease. And pre-load the diff slightly tighter than your bench-test sweet spot so it lands on target at race temperature, not at room temp.
Don't mix. Carbide balls (HRA 90+) on steel plates (HRC 58-62) will groove the plates within a few heats because the ball hardness is roughly double the plate hardness. You'll get short-term performance gain followed by accelerated plate wear and a notchy feel.
Run steel balls on steel plates, or carbide balls on tungsten carbide plates. Both pairings wear at compatible rates. The tungsten plate plus carbide ball combination is what most race-spec kits ship as standard now, and the plates routinely last 50+ heats versus 10-15 for the mismatched setup.
References & Further Reading
- Wikipedia contributors. Ball differential. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
