A Conical Roller Thrust Bearing is an axial-load bearing built from a ring of tapered (conical) rollers running between two flat or grooved races so the roller axes converge on the shaft centreline. Unlike a flat thrust washer or a ball thrust bearing, the conical geometry gives true rolling contact with no scrubbing across the race. It exists to carry heavy axial thrust at moderate speeds — vertical mill spindles, screw-down jacks on rolling mills, ship reduction gears — where a ball thrust would brinell within hours. A single SKF 29420 E spherical/conical thrust unit handles 2,400 kN static axial load.
The Conical Roller Thrust Bearing in Action
The trick is in the cone angle. If you put cylindrical rollers between two flat plates and rotate one plate, every roller scrubs �� the inner end of the roller travels slower than the outer end, but a cylinder forces them to travel the same. That scrubbing tears races apart inside 200 hours under load. A Conical Roller Thrust Bearing solves this by tapering the roller so its inner end is smaller in diameter than its outer end, with the cone apex hitting the shaft axis. Now every point along the roller has a surface speed exactly matching its race contact radius. True rolling. No scrub.
The roller axes sit at an angle — typically 5° to 15° off perpendicular to the shaft for a pure tapered roller thrust bearing, steeper for spherical roller thrust variants pushed to 45°. That angle decides the split between axial load capacity and the small radial component the bearing can also tolerate. A cage holds the rollers at fixed circumferential spacing so they don't crowd or skew, and the races are usually case-hardened 100Cr6 (52100) bearing steel ground to a surface finish below Ra 0.2 µm. If the race finish drifts above Ra 0.4 µm you'll see line-contact spalling on the loaded race within the first 500 hours, because the oil film can't separate the roller from the asperity peaks.
Get the preload wrong and the bearing eats itself. Too loose and the rollers skid instead of rolling — skidding generates a smear pattern on the race and welds metal to roller ends. Too tight and contact stress climbs above 4,000 MPa, fatigue life collapses, and you'll measure rising temperature within 30 minutes of startup. The design sweet spot is a light axial preload that just removes endplay, typically 0.02 to 0.05 mm of measured shaft movement under a dial indicator. The other failure you see in the field is roller skew — when the cage wears or a roller end gets damaged, the roller cocks in its pocket and starts grinding the race flange. That shows up as fine bronze powder in the oil if the cage is brass, or as a sudden 5-10°C jump in bearing housing temperature.
Key Components
- Tapered Rollers: Hardened 100Cr6 steel rollers with a cone half-angle typically 2° to 8°, ground to a diameter tolerance of ±2 µm within a single set. The cone apex must intersect the shaft axis exactly — a 0.1° error introduces sliding and halves the L10 life.
- Shaft Race (Inner Race / Cone Washer): The flat or shallow-grooved washer that bolts or presses against the shaft shoulder. Surface hardness 58-62 HRC, finish below Ra 0.2 µm. It carries the load from the shaft into the rollers.
- Housing Race (Outer Race / Cup Washer): Sits in the housing bore against a machined shoulder. Same hardness and finish spec as the shaft race. Squareness to the bore axis must be within 0.025 mm or the rollers load unevenly across the ring.
- Cage (Roller Retainer): Pressed steel, machined brass, or polyamide depending on speed. Holds rollers at equal circumferential spacing and prevents skew. Brass cages handle higher temperatures (up to 200°C); polyamide caps out around 120°C but runs quieter.
- Lubrication Path: Either oil bath, oil mist, or grease. For mill-duty bearings running above 50 RPM, circulating oil at 40-60 cSt viscosity is standard — grease packs out and starves the rolling contact above that speed.
Real-World Applications of the Conical Roller Thrust Bearing
Conical Roller Thrust Bearings show up wherever axial load is heavy, shock-loaded, or sustained for hours. The common thread is a vertical shaft or a horizontal shaft with serious thrust component — gearbox output shafts, mill screw-downs, vertical pump impellers, ship reduction gears. They compete with hydrostatic thrust pads at the very high-load end and with spherical roller thrust bearings where misalignment is a concern.
- Steel Rolling Mills: Screw-down mechanisms on a 4-high hot strip mill — Davy or SMS designs use double-row tapered roller thrust bearings under each screw to carry the rolling separation force, often 30-50 MN per stand.
- Marine Propulsion: Main thrust block on a ship reduction gearbox — Wärtsilä and MAN propulsion gear sets use a Michell-style thrust pad bearing combined with conical roller thrust units on the intermediate pinion shafts.
- Vertical Pumps: Sulzer and Flowserve vertical turbine pumps for power-station cooling water use a conical roller thrust bearing at the top of the line shaft to carry the full hydraulic down-thrust of the impeller stack, typically 50-200 kN.
- Wind Turbine Gearboxes: ZF and Winergy main gearbox planet carriers run tapered roller thrust units on the high-speed pinion to absorb helical-gear axial reaction force, sized for 175,000-hour design life.
- Heavy Press Machinery: Schuler mechanical presses use conical roller thrust bearings on the eccentric shaft end caps to handle the cyclic axial component from the helical drive gear.
- Mining Mill Drives: FLSmidth and Metso SAG mill girth gear pinions ride on tapered roller thrust assemblies to carry helical-gear thrust at 12-15 RPM under 20+ MW input power.
The Formula Behind the Conical Roller Thrust Bearing
The number you actually care about is bearing life — how many hours the thing runs before fatigue spalling forces a replacement. The L10 life equation tells you what fraction of bearings will survive a given duty before 10% of them fail. What matters in practice is how brutally life scales with load: doubling the axial load doesn't halve the life, it cuts life by a factor of about 10 for roller bearings. Run a bearing at half its rated capacity and you get tens of thousands of hours. Run it at rated capacity and you might see 1,500 hours. Push it to 150% and you're replacing it in a long weekend. The sweet spot for mill-duty service is sizing the bearing so equivalent dynamic load P sits between 25% and 40% of dynamic capacity C — that puts L10 above 30,000 hours with margin for shock loads.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| L10h | Basic rating life — hours until 10% of a population fails by fatigue | hours | hours |
| n | Rotational speed | RPM | RPM |
| C | Basic dynamic load rating from the bearing catalogue | N (or kN) | lbf |
| P | Equivalent dynamic axial load on the bearing | N (or kN) | lbf |
| 10/3 | Life exponent for roller bearings (3 for ball bearings) | dimensionless | dimensionless |
Conical Roller Thrust Bearing Interactive Calculator
Vary bearing radii and shaft speed to see true-rolling surface speeds, velocity ratio, and the scrub that a cylindrical roller would create.
Equation Used
The calculator applies the article relationship that surface speed is proportional to race contact radius. A conical roller avoids scrubbing when its taper makes the outer-end speed V2 and inner-end speed V1 match V = omega r at each radius.
- Race angular speed is uniform across the contact path.
- Outer radius is constrained to be at least 1 mm larger than inner radius in the calculation.
- Scrub index compares the end-speed spread with the average end speed.
Worked Example: Conical Roller Thrust Bearing in a vertical CNC turret lathe spindle
A machine-tool rebuild shop in Hartford Connecticut is refitting the vertical thrust bearing on a Bullard 56-inch vertical turret lathe spindle that carries a chuck and workpiece weighing 8,500 kg. They have specced an SKF 29328 E spherical roller thrust bearing with C = 1,290 kN dynamic capacity. The spindle runs at a nominal 60 RPM during heavy turning, with a low-end creep speed of 15 RPM for boring deep bores and a high-end finishing speed of 150 RPM. They need to know expected L10 life across that range to plan rebuild intervals.
Given
- Workpiece + chuck mass = 8,500 kg
- Axial load P (= mass × g) = 83.4 kN
- Bearing dynamic capacity C = 1,290 kN
- nlow / nnom / nhigh = 15 / 60 / 150 RPM
- Life exponent (roller bearing) = 10/3 —
Solution
Step 1 — compute the load ratio C/P. This is the headroom the bearing has over the actual axial load:
That is generously sized — P sits at about 6.5% of C, which is well below the 25-40% sweet spot. The bearing is going to last a very long time at this load. Now raise that ratio to the 10/3 power, which is where the brutal scaling lives:
Step 2 — at nominal 60 RPM, plug into the L10 hours formula:
That number is absurd in the real world — no spindle bearing actually lives 250 years. What it tells you is that fatigue is not the failure mode here. Lubrication breakdown, contamination, or seal wear will kill this bearing long before fatigue does. Realistic service life will be 80,000-120,000 hours governed by oil cleanliness and seal condition.
Step 3 — at the low end, 15 RPM creep speed, life scales inversely with speed:
At 15 RPM you're firmly in the regime where oil-film thickness becomes marginal — the elastohydrodynamic film needs surface speed to build up, and below about 20 RPM you risk metal-to-metal contact at the roller-race interface. Life on paper goes up but real-world wear from boundary lubrication goes up too.
Step 4 — at the high end, 150 RPM:
Still enormous. At 150 RPM you're nowhere near the bearing's speed limit (the SKF 29328 E is rated for around 1,300 RPM with oil), so heat is not an issue. The practical lifespan is still set by the oil and seals, not the rolling elements.
Result
Nominal L10 life at 60 RPM works out to roughly 2. 2 million hours — a number that means fatigue is effectively a non-issue and the bearing is heavily oversized for this duty. Across the operating range, life ranges from 8.8 million hours at 15 RPM creep to 880,000 hours at 150 RPM finishing speed; the sweet spot for actually loading the bearing properly would be a workpiece three to four times heavier, which is what Bullard sized this spindle for in the first place. If you measure bearing temperature climbing above 70°C in service despite the calculation predicting infinite life, the failure mode is not fatigue — look for oil viscosity dropped below 32 cSt at operating temperature (wrong oil grade), seal lip drag from a hardened Viton seal that has lost elasticity, or roller skew from a cracked brass cage segment letting one roller cock in its pocket. Any of those will show up as housing temperature rise of 5-15°C above baseline within the first hour of operation.
Choosing the Conical Roller Thrust Bearing: Pros and Cons
Picking a thrust bearing comes down to load magnitude, speed, misalignment tolerance, and budget. Conical roller thrust bearings sit in the heavy-load, moderate-speed band. Below them in load capacity is the ball thrust bearing — cheap and fast but it brinells under shock. Above them in capacity for the same envelope is the spherical roller thrust bearing, which adds misalignment tolerance at higher cost.
| Property | Conical Roller Thrust Bearing | Ball Thrust Bearing | Spherical Roller Thrust Bearing |
|---|---|---|---|
| Axial load capacity (typical 200 mm bore) | 800-2,000 kN dynamic | 150-400 kN dynamic | 1,500-4,000 kN dynamic |
| Maximum speed (oil-lubricated) | Moderate, 1,000-2,500 RPM | High, 3,000-6,000 RPM | Moderate, 800-2,000 RPM |
| Misalignment tolerance | Poor, <0.05° before edge loading | Poor, <0.1° | Excellent, up to 2-3° |
| Shock load tolerance | Good — line contact spreads peak stress | Poor — point contact brinells | Excellent |
| Relative cost (200 mm bore) | Mid — roughly 1.5× ball thrust | Low baseline | High — 2-3× ball thrust |
| L10 life at 30% rated load | ~30,000-50,000 hours | ~15,000-25,000 hours | ~40,000-60,000 hours |
| Typical application fit | Mill screw-downs, vertical lathes, gear thrust | Light vertical pumps, machine-tool feed screws | Wind turbines, marine drives, large mining mills |
Frequently Asked Questions About Conical Roller Thrust Bearing
Heat in a thrust bearing almost never comes from fatigue — it comes from churning losses, wrong oil viscosity, or excessive preload. If you've used the calculated life as your only check, you've missed the dynamics. At low axial loads relative to capacity (under about 10% of C), the rollers can skid instead of pure-rolling, and skidding generates more heat than rolling would. The fix is either reduce oil level so the rollers aren't churning a sump full of oil, switch to a lighter viscosity grade matched to operating temperature, or add a small spring preload to keep rollers loaded enough to roll cleanly.
Quick check: measure housing temperature 30 minutes after startup with the design oil. If it stabilises above 65°C with ambient around 20°C, your operating temperature is wrong for the oil grade in the housing.
The deciding factor is shaft alignment, not load. A vertical pump line shaft 6-12 metres long bolted through multiple intermediate guide bearings will inevitably accumulate angular misalignment at the thrust bearing — typically 0.1° to 0.5° depending on installation precision. A pure conical (tapered) roller thrust bearing edge-loads above 0.05° misalignment and you'll see localised spalling on the race within 2,000 hours.
A spherical roller thrust bearing tolerates 2-3° of misalignment because the outer race is spherically ground. For any line-shaft pump over about 4 metres long, default to spherical. Reserve true tapered-roller thrust units for short, rigid shafts where you can dial-indicator the squareness within 0.025 mm.
You're loose by enough that the rollers are skidding instead of rolling. Endplay above about 0.10 mm in a thrust bearing means the axial load is bouncing the rollers off their loaded race and back, and during the off-load portion of each revolution the rollers don't have enough traction to keep rolling — they skid. Skidding shows up as a polished smear band on one race and bronze flecks in the oil if the cage is brass.
The cause is usually a missed shim during assembly, a crushed gasket under the housing cap, or a shaft shoulder that wasn't faced flat. Pull the cap, check shoulder squareness with a dial indicator (target under 0.025 mm runout), and re-shim to land in the 0.02-0.05 mm preload window.
Not as a single unit — a single conical roller thrust bearing only carries load in one axial direction. The roller cone apex points one way, and reversing the load lifts the rollers off their race entirely. For reversing or alternating axial loads you need a double-direction arrangement: two opposed thrust bearings, or a purpose-built double-row unit like an SKF 234400 series double-direction angular contact thrust bearing.
Schuler and Komatsu mechanical-press eccentric shafts use exactly that arrangement — a back-to-back pair preloaded against each other through a clamping nut, so whichever direction the load reverses to, one bearing is always engaged.
Dynamic capacity C is the load that gives 1 million revolutions of L10 fatigue life — it's a fatigue-driven number tied to rolling contact. Static capacity C₀ is the load that produces 0.0001 × roller-diameter of permanent indentation in the race — it's a yield-driven number for stationary or slow-oscillating loads.
If your shaft turns continuously, size on C and the L10 equation. If your shaft sits stationary under load most of the time and only rotates occasionally — like a slewing bearing on a crane or a valve actuator stem — size on C₀, because fatigue won't accumulate but a single overload event can brinell the race permanently. Mill screw-downs need both checked: C₀ for the stalled rolling-force peak, C for the cumulative duty.
Yes, and the temperature being normal is misleading. Fine metallic particles in oil from a thrust bearing are usually one of three things: cage wear (brass flakes, bronze colour), race micropitting (silvery grey iron particles under 10 µm), or roller-end scuffing (larger steel slivers). All three precede catastrophic failure by hundreds to a few thousand hours, and none of them produces a temperature signature until the damage is already advanced.
Get a particle count and ferrography done. A jump from ISO 18/16/13 cleanliness to 21/19/16 with a rising ferrous-particle trend means the bearing is degrading even if the thermocouple says everything is fine. Replace the oil, clean the housing, and plan a bearing inspection at the next available shutdown.
References & Further Reading
- Wikipedia contributors. Tapered roller bearing. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
