A full complement ball bearing, which packs the maximum number of balls by omitting the cage, trades speed for radial capacity, and the geometry that makes it strong radially leaves it unable to direct axial force into the raceways [S1].
In practical terms, a 6203 full complement ball bearing reaches 14,000 N dynamic and 7,000 N static load ratings, about 45% above a caged 6203ZZ at 9,550 N and 4,800 N, but the limit speed collapses from 17,100 rpm (grease) to 2,000 rpm (grease) and from 20,100 rpm (oil) to 3,000 rpm (oil) [S1]. That speed loss is the mechanical price of ball-on-ball friction, and it is also the reason thrust load fails.
Contact Geometry Decides What a Bearing Can Carry
Ball bearings transfer load through point contact, roller bearings through line contact, and that single distinction sets the load-direction rules engineers have to follow at the bench [S3]. A radial deep-groove ball bearing has groove radii matched to the ball so the contact ellipse sits roughly perpendicular to the shaft axis; any axial component is reacted only by the small shoulder lip on each race, which is why deep-groove types are rated for some axial load, not much.
Thrust bearings flip that geometry: the balls run in a groove angle close to 90 degrees relative to the shaft, so the contact ellipse is aligned to take force along the axis. The two designs are not interchangeable, and the standard warning in supplier catalogs is that a thrust bearing should not be subjected to a radial load at all because the raceways have no curvature to react sideways force [S2].
What Removing the Cage Actually Changes
The cage in a conventional bearing does three jobs at once: it spaces the rolling elements evenly around the circumference, holds them apart so they do not collide, and (critically) keeps every ball following the same contact-angle path. Drop the cage, as a full complement design does, and the load capacity rises because more balls share the raceway: dynamic and static load ratings both climb by up to 45% in the 6203 example, and a 608 full complement variant follows the same pattern [S1].
What is lost is angular control. Without a cage pocket, each ball drifts in its groove under load, and under any axial component the balls rotate at slightly different speeds because of the helix of the contact. With nothing to keep them apart, the adjacent balls rub shoulders, friction jumps, lubricant films break down, and the bearing either seizes or runs catastrophically hot. The same mechanism that limits full complement designs to roughly one-tenth the speed of a caged equivalent is the mechanism that destroys them under thrust [S1].
Why the Raceway Itself Pushes Back
A standard deep-groove ball bearing has a groove radius only slightly larger than the ball radius, typically about 52% of the ball diameter, which gives the contact ellipse room to tilt a few degrees when axial force appears. That tilt is what converts a small axial load into normal force at the shoulder lip. In a full complement bearing the balls are packed to the maximum fill, so any axial shift forces the ball at the contact point to climb the shoulder and wedge its neighbor, rather than rolling cleanly into the lip. [S1]
For a true thrust bearing, the washers are flat or have a flat-bottomed groove with a 90-degree contact angle, so axial force pushes the ball straight into the hardened race with no sideways component to manage. Trying to pull axial force through a full complement radial design is asking the contact ellipse to do two things at once: carry the rated radial load and steer a sideways force, with no cage to keep the ball train on path [S2].
When the Spec Actually Fits (and When It Does Not)
Full complement ball bearings are a strong choice for slow, heavily loaded radial applications: crane sheaves, conveyor idlers, winch drums, kiln trunnions, and slow gearbox support positions where the rated 2,000 to 3,000 rpm grease or oil limit is comfortably above the operating speed [S1]. The designer gains roughly 45% radial capacity and accepts a higher friction torque in return, plus the obligation to specify a relubrication interval much shorter than a caged equivalent.
They are the wrong choice anywhere thrust matters, anywhere speed exceeds about 3,000 rpm, and anywhere misalignment exceeds the very small clearance the design tolerates. For those cases the spec should move to a caged deep-groove ball bearing (some axial capacity, high speed), an angular contact bearing pair (DB or DF mount, defined contact angle, real axial rating), a cylindrical roller bearing with an L-shaped flange (NCF or NJG series, true semi-locating function for one-direction thrust), or a purpose-built thrust bearing if the load is purely axial [S5].
Comparison: Where Each Bearing Type Actually Wins
On the four criteria that drive selection, full complement ball, caged deep-groove ball, cylindrical roller, and thrust ball sort out as follows. Radial capacity: full complement ball highest in the ball family, cylindrical roller higher still on line contact. Axial capacity: thrust ball highest by design, angular contact next, full complement ball effectively zero. Maximum speed: caged deep-groove ball wins, with full complement ball limited to roughly 2,000 to 3,000 rpm depending on lubricant [S1]. Friction and heat: full complement ball worst because of inter-ball rubbing, cylindrical roller intermediate, caged ball and thrust ball lowest.
The 6203 family captures the trade cleanly: 6203ZZ delivers 9,550 N dynamic at 17,100 rpm (grease) with low friction, while the 6203 full complement version delivers 14,000 N dynamic at only 2,000 rpm (grease), a 1.47x load gain in exchange for an 8.5x speed penalty [S1]. Designers reading that number know the bearing belongs on a slow shaft, and that any axial force on that shaft is a different problem entirely.
Failure Modes and Field Signals
The classic full complement ball failure under unintended thrust starts with a hot spot on one shoulder lip, progresses to discoloration and lubricant carbonization in that arc, and ends with the balls welding to the raceway over a sector of about 30 to 60 degrees. Inspection will show the lubricant blackened on the loaded side, the cage (where fitted) smeared sideways, and the ball complement with witness marks where neighbor contact occurred. None of those signals appear in a clean radial-only application. [S1]
Engineers specifying full complement designs should pair the bearing choice with a documented thrust source check: gear tooth forces, belt tension vectors, thermal shaft growth, and any external axial load on the shaft must be summed, and if the sum is non-zero the bearing arrangement needs an axial-reacting element elsewhere in the stack, not a full complement radial. A matched pair of angular contact bearings, or a cylindrical roller with an L-flange, handles the axial path without ruining the full complement's radial advantage [S5].
Sourcing and Specification Discipline
Published ratings for full complement ball bearings come from manufacturer catalogs and product tables, and the underlying methodology is the ISO 281 dynamic load rating and ISO 76 static load rating framework that the whole rolling-element industry uses [S1]. The 45% capacity uplift cited for the 6203 full complement versus the 6203ZZ is a like-for-like catalog comparison under those ratings, and the speed limits are the catalog grease and oil limits, not field-corrected values.
For sourcing, the same suppliers that publish caged bearing catalogs (SKF, NTN, NSK, FAG, Schaeffler/INA, Timken) all carry full complement ball lines, and the ball bearing reference page lays out the dimensional series that the full complement variants slot into. Cross-check the exact part number in the manufacturer's table before assuming a 45% uplift transfers to a different bore size, because the gain depends on the fraction of the ball complement the cage was occupying, and that fraction varies with internal geometry.
Next signal to track: a published catalog comparison of full complement ball versus full complement cylindrical roller (NCF, NJG) on identical boundary dimensions, which would let a designer pick between line-contact radial capacity and point-contact radial capacity with the same axial story [S5]. A second trackable item is any ISO or ABMA revision clarifying how the 45% capacity figure is normalized across manufacturers, because the number is widely repeated but the test conditions are not standardized across catalogs.
The underlying component specifications are covered under electronic load.
For related coverage, see Spherical Plain vs Roller Bearing on an Oscillating Pivot.