A cone-on-cone rolling pair is a higher kinematic pair in which two cones contact along a common straight line element and transmit motion through pure rolling, with their apexes meeting at a single point. You will find it in traction drives, paper-mill calender adjustments, and variable-speed friction transmissions where shafts cross at an angle and gear teeth would be too noisy or too coarse. The pair converts one shaft's rotation into another's at a ratio set by the cone half-angles, with no sliding along the contact line. The outcome is silent, backlash-free angular transmission between intersecting shafts.
Cone-on-cone Rolling Pair Interactive Calculator
Vary cone half-angles, input speed, and apex error to see the pure-rolling speed ratio, output rpm, shaft angle, and misalignment risk.
Equation Used
The calculator applies the article pure-rolling condition: the tangential surface speeds match along the shared cone contact line, so omega_A sin(alpha) equals omega_B sin(beta). The shaft crossing angle is Sigma = alpha + beta. Apex error is also shown as e/L because the article notes that even a small apex miss can turn rolling into scrubbing.
FIRGELLI Automations - Interactive Mechanism Calculators.
- Cone apexes are intended to coincide for pure rolling.
- alpha and beta are cone half-angles in degrees.
- Cone A is the input cone and cone B is the driven cone.
- Apex error is reported as offset divided by cone length.
How the Cone-on-cone Rolling Pair Actually Works
Two cones touch along a straight line — not a point, not a circle — and that line passes through both apexes. Reuleaux classified this as a higher element pair because the contact is a line rather than a surface. For pure rolling to actually happen, the apexes must coincide exactly. Miss that coincidence by even 0.5 mm in a 100 mm cone and you stop rolling and start scrubbing — the contact line skews, one end of the line slips forward while the other slips back, and you get heat, wear streaks, and a measurable drop in transmitted torque.
The speed ratio falls out of the geometry. At any radius r along the contact line, the surface speed of cone A must equal the surface speed of cone B, otherwise the cones slide. That gives ωA × sin(α) = ωB × sin(β), where α and β are the half-angles of the two cones. The half-angles also fix the shaft angle: Σ = α + β. So once you decide the shaft crossing angle and the ratio you want, the cone geometry is locked.
Where does it fail? Three places. Apex misalignment as already mentioned — this is the killer in field-built traction drives. Insufficient normal load at the contact line, which lets the cones micro-slip under torque (Hertzian contact theory says you need a preload that scales with the tangential load you want to transmit, typically 3-5× the tangential force for a steel-on-steel traction drive). And surface contamination — a film of cutting oil on a dry friction cone drops the traction coefficient from around 0.08 to under 0.02, and the drive simply spins. If you are designing this pair, the bore concentricity on the cone shafts must be inside about 0.02 mm TIR or the apex wanders as the shafts turn, and you get a 1-per-rev torque pulse the operator will feel through the handwheel.
Key Components
- Driving cone: The input cone, fixed to the input shaft. Its half-angle α and the shaft crossing angle Σ together determine the ratio. Surface finish is critical — Ra below 0.4 µm for hardened steel traction cones, otherwise the asperities tear up the mating surface within a few hundred hours.
- Driven cone: The output cone with half-angle β. Its apex must lie on the same point in space as the driving cone's apex. The cone material is usually paired dissimilarly — hardened steel against a phenolic or polyurethane-faced cone — to keep the traction coefficient predictable around 0.05-0.10.
- Contact line element: The straight line where the two cones touch. It runs from the shared apex outward and rotates around it. Hertzian line-contact stress on this element governs life — keep peak contact stress below about 1,200 MPa for through-hardened bearing steel cones to hit 10,000 hours.
- Preload mechanism: A spring, screw, or hydraulic piston that pushes the cones together along the contact normal. The preload must equal roughly tangential force divided by the traction coefficient, with a 1.5× safety margin. Too little preload and the drive slips; too much and the cones brinell at the contact line.
- Shaft bearings: Both shafts need bearings that take radial load AND the axial preload reaction. Angular-contact ball bearings or tapered rollers are typical. Bearing axial slop above 0.01 mm shows up directly as apex wander and ratio variation.
Who Uses the Cone-on-cone Rolling Pair
The cone-on-cone rolling pair shows up wherever you need quiet, backlash-free transmission between intersecting shafts at a fixed ratio, or where you want a continuously variable ratio by sliding one cone axially against another. It is not a high-power solution — gears beat it on torque density every time — but for low-noise, low-vibration, or precision-positioning duties it earns its place. Failure in service is almost always traceable to apex misalignment, contamination, or under-preload, in that order.
- Paper and textile machinery: Calender roll speed-matching drives on Voith and Valmet paper machines, where two cones in light contact synchronize roll surface speeds without the torque ripple of a geared coupling.
- Precision instruments: Theodolite and surveying-instrument fine-motion drives — Wild Heerbrugg and Kern instruments used cone-on-cone pairs in their slow-motion tangent screws for backlash-free angular adjustment.
- Variable-speed transmissions: Graham Variable Speed Transmissions and the Kopp variator both use a cone-on-cone or cone-on-disc principle for stepless ratio change in industrial mixers and conveyors.
- Bevel friction drives: Old machine-tool feed drives — the South Bend lathe apron used a friction-cone variant for the cross-feed selector before all-gear apron designs took over in the 1940s.
- Robotics demonstrators: University kinematics labs at Cornell and Cambridge build cone-on-cone rigs from the Reuleaux model collection to demonstrate higher-pair rolling contact to mechanical engineering students.
- Phonograph and clockwork: Edison cylinder phonograph governors used cone friction pairs to transmit motion from the spring barrel to the speed regulator without introducing audible gear chatter into the pickup.
The Formula Behind the Cone-on-cone Rolling Pair
The speed ratio of a cone-on-cone rolling pair is set entirely by the cone half-angles. At the low end of the typical operating range — shaft crossing angles around 30° — you get gentle ratios near 1:1 with shallow cones that are easy to align but transmit limited torque per unit length of contact line. At the nominal 90° crossing (perpendicular shafts) you hit the design sweet spot where the geometry is symmetrical and the preload reaction sits cleanly on the bearings. Push the crossing angle past 150° and you are running near-flat cones that behave more like discs, with apex coincidence becoming hypersensitive to thermal growth. The formula below tells you the output speed at a given input speed for any cone-pair geometry.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| ωA | Angular velocity of driving cone | rad/s | RPM |
| ωB | Angular velocity of driven cone | rad/s | RPM |
| α | Half-angle of driving cone | degrees or radians | degrees |
| β | Half-angle of driven cone | degrees or radians | degrees |
| Σ | Shaft crossing angle (Σ = α + β) | degrees or radians | degrees |
Worked Example: Cone-on-cone Rolling Pair in a paper-mill calender speed-trim drive
You are sizing a cone-on-cone trim drive between two calender roll shafts that cross at 90°. The driving cone has a half-angle of α = 30°, leaving the driven cone at β = 60°. Input speed from the trim motor is 300 RPM nominal, with an operating window of 150 RPM at slow-thread-up to 600 RPM at full production line speed. You need the output speed at each operating point to confirm the downstream roll keeps surface-speed match.
Given
- α = 30 degrees
- β = 60 degrees
- ωA,nom = 300 RPM
- ωA,low = 150 RPM
- ωA,high = 600 RPM
Solution
Step 1 — confirm the crossing angle is consistent: Σ = α + β = 30° + 60° = 90°. Geometry checks out.
Step 2 — at nominal 300 RPM input, apply the ratio:
That's the design point. At 173 RPM the driven shaft turns smoothly with the contact line carrying maybe 200 N of preload against perhaps 80 N tangential — comfortably inside the 0.08 traction-coefficient envelope for hardened steel against polyurethane.
Step 3 — at the low end of the typical operating range, 150 RPM input during thread-up:
At this speed the cones barely turn — you can watch the contact line walk if the apex is even slightly off. This is also where you'll catch alignment faults during commissioning, because the operator can feel torque ripple through the trim handwheel that disappears at higher speeds.
Step 4 — at the high end, 600 RPM at full line speed:
In theory clean rolling continues, but in practice above about 500 RPM you start seeing centrifugal lift on the larger cone face, the contact patch narrows, and the effective traction coefficient drops 15-20%. If you don't bump preload by a similar amount, the cone slips and the calender goes out of speed-match.
Result
Nominal output is 173. 2 RPM at 300 RPM input — the 1:√3 ratio falls straight out of the 30°/60° cone geometry. Across the operating range, the driven cone runs from 86.6 RPM at thread-up to 346.4 RPM at full line speed, with the cleanest behaviour clustered around the 150-300 RPM nominal band where preload, traction coefficient, and contact-line stress all sit in their design windows. If you measure 160 RPM instead of 173 RPM at nominal input, suspect micro-slip from low preload first — back-calculate required normal force from your tangential torque and the traction coefficient, and you usually find the spring has settled 10-15%. If the output speed wanders by ±2-3 RPM cycle-to-cycle, the apex is offset — measure both shaft positions against the theoretical apex coordinate, you are almost always 0.1 mm or more off. If you see scoring on the polyurethane cone face after 200 hours, the steel cone surface finish is rougher than 0.4 µm Ra and is acting like a file.
Cone-on-cone Rolling Pair vs Alternatives
Cone-on-cone rolling pairs compete against bevel gears and against cone-on-disc traction drives for intersecting-shaft duty. The trade is roughly: gears for torque density, cones for quietness and backlash, cone-on-disc for ratio variability. Pick on the dimension that hurts most in your application.
| Property | Cone-on-cone rolling pair | Bevel gear pair | Cone-on-disc traction drive |
|---|---|---|---|
| Torque capacity (typical, 100 mm pitch) | 50-200 Nm | 500-3000 Nm | 30-150 Nm |
| Backlash | Zero (rolling contact) | 0.05-0.2 mm at pitch line | Zero |
| Noise level at 1500 RPM | 55-65 dB | 75-90 dB | 55-65 dB |
| Speed ratio range | Fixed by α, β | Fixed by tooth count | Continuously variable, typically 4:1 |
| Setup tolerance (apex/centre) | ±0.02 mm TIR critical | ±0.05 mm acceptable | ±0.05 mm acceptable |
| Service life under steady load | 8,000-15,000 hr | 20,000-50,000 hr | 5,000-10,000 hr |
| Cost (machined pair, low volume) | $$$ — precision grinding | $$ — standard tooling | $$$$ — variator hardware |
| Best application fit | Quiet, backlash-free fixed ratio | High-torque power transmission | Variable-speed industrial drives |
Frequently Asked Questions About Cone-on-cone Rolling Pair
Heat at the contact line nearly always means micro-slip, not bearing or windage loss. If your apex coincidence is off by even 0.1 mm in a 100 mm cone, the contact line carries differential surface velocity along its length — one end rolls faster than the other, and the difference dissipates as friction heat. Calculated power loss assumes pure rolling, which you do not have.
Diagnostic check: dust the cone faces with engineer's blue, run for 30 seconds at low speed, and inspect. A clean diagonal wipe pattern means rolling. A smeared or zig-zag pattern means slip. Re-shim the bearing pedestals until the apex coordinates match within 0.02 mm and the heat usually disappears.
At 200 Nm you are sitting right at the upper edge of what a cone-on-cone pair handles comfortably, and at maybe 10% of what a bevel gear pair does without breathing hard. Pick cones only if you have a hard requirement that gears cannot meet — usually noise below 65 dB, zero backlash for positioning, or a specific need for slip protection as a torque limiter.
If your duty is ordinary power transmission, bevel gears win on cost, life, and torque margin. The cone-on-cone pair earns its place in instruments, calenders, and slow precision drives — not in general-purpose gearboxes.
Two common causes. First, the traction coefficient you used was probably a clean-and-dry catalog figure around 0.08-0.10. Real machinery has airborne oil mist, and a thin film drops effective μ to 0.02-0.04. Re-size preload assuming the lower value, or seal the contact zone.
Second, preload springs settle. A coil spring loses 10-15% of its installed force in the first 500 hours, and a Belleville stack can lose more if it sees thermal cycling. If you sized at exactly the calculated preload with no margin, you are under-preloaded by month two. Always design with at least a 1.5× safety factor on preload.
Yes, the geometry works for any Σ between roughly 10° and 170°. Below 10° the cones get pencil-thin and contact-line stress shoots up. Above 170° they flatten into near-discs and apex coincidence becomes hypersensitive to thermal growth — a 50°C temperature rise can shift apex position by 0.1 mm on a 200 mm cone, enough to lose rolling.
The sweet spot for industrial use is 60° to 120°. Outside that band, evaluate cone-on-disc or angular-bevel-gear alternatives before committing.
That magnitude of error is almost always cone half-angle deviation, not slip. If your driving cone was ground to 30.5° instead of 30.0°, sin(30.5°)/sin(60°) gives you 1.762 — already most of your error. Verify the actual half-angles with a sine bar or a CMM trace before assuming anything else.
If the half-angles are correct on inspection, check thermal expansion at running temperature. A cone running 40°C above ambient grows enough to shift the effective contact radius by a fraction of a percent, and that ratio change is real, not measurement noise. The fix is to spec your input speed at operating temperature, not cold.
For steel-on-steel traction cones, use through-hardened bearing steel (52100 or equivalent) at 60-62 HRC on both members, with surface finish at Ra 0.2-0.4 µm. Peak Hertzian contact stress must stay below about 1,200 MPa to hit 10,000 hours — anything above that and you will see subsurface fatigue spalling at maybe 3,000-5,000 hours.
If one cone is steel and the other is faced with polyurethane or phenolic, the polymer is the wear member. Spec it as replaceable — typical life is 4,000-8,000 hours depending on contact pressure, and you want the steel cone to outlive three or four polymer cones, not the other way round.
References & Further Reading
- Wikipedia contributors. Kinematic pair. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
