Caster Bearing Types Compared: 5 Checks Before You Specify Ball, Roller or Plain

Rule out bearings that cannot carry the load

Load per wheel is a hard gate, not a preference. One source states that above 1,500 lb per wheel, roller bearings (caged needle or tapered cone) or double-row ball bearings are required, and that roller bearings resist shock loads better than any ball bearing¹⁴.

A buyer who compares push force first will re-specify the caster later, because the ball option is not a cheaper compromise above that line — it is unqualified. The same source gives annular ball bearings a load range of 300-1,500 lb per wheel depending on bearing size and axle¹² (supplier-reported figure).

So the two ranges overlap only up to 1,500 lb, and the overlap is where the push-force and cost trade actually happens. Walk a 2,000 lb per wheel towed train: the 1,500 lb threshold excludes ball bearings outright and forces caged needle or tapered cone roller, with double-row ball as the alternative¹⁴.

That is a construction change, not a bearing swap, and it usually changes the hub and axle too. Note the sources do not state which component — bearing, wheel or bracket — becomes limiting at that load, so treat the threshold as a selection gate rather than a failure prediction.

Before you quote anything, write the load per wheel on the drawing and check it against the 1,500 lb line.

Is push force set by the bearing or the wheel?

Push force is the sum of starting resistance and rolling resistance. Bearing type sets the rolling part; tread hardness and wheel diameter set the rest.

One source reports ball bearing rolling resistance at roughly 40% of a plain bearing's effort at the same load¹², and a plain-to-precision-ball swap drops push force 30-40%¹⁷. A third reports the same swap at roughly 40%¹⁴ (supplier-reported figure).

Those three figures describe one change, and they are close enough to plan around. The wheel side is just as strong a lever.

One source reports each extra inch of wheel diameter cuts push force 10-15%, and dropping durometer from 95A to 85A polyurethane cuts it 15-20% with reduced capacity¹⁷. Harder tread also lowers starting resistance, which matters most as load rises⁴.

So a bearing upgrade and a wheel change are partly substitutable on the same ergonomic target. For a cart that needs one operator instead of two, price both routes before you commit.

The sources do not give absolute push force in newtons or a coefficient of rolling resistance per bearing type, so you cannot quote a target figure — you can only compare the percentage levers. If your operator complaint is breakaway force rather than sustained effort, ask the supplier which lever they are quoting.

Match the lubrication regime to who greases it

An unsealed plain or roller bearing only reaches its rated life if someone greases it on schedule. One source states plain and roller bearings with grease zerks need lubrication every 30-90 days in normal industrial duty, and weekly under heavy towing¹⁴.

Sealed ball and delrin bearings are lubricated for life and never take external grease¹⁴. That is the whole trade: the sealed unit costs more up front and removes the schedule.

The same source claims sealed precision bearings cut rolling resistance 10-15% versus standard annular ball and extend service life 5-10x where operators cannot or will not follow a grease schedule¹⁴ (supplier-reported). Treat that 5-10x as a supplier claim, not a measured life (supplier-reported).

For a continuously running AGV, the arithmetic is simple: a plain or roller bearing needs 4-12 grease events a year at the 30-90 day interval, weekly under towing, and a sealed unit needs none¹⁴.

A typical ABEC 1 annular ball bearing lasts 2-4 million revolutions under rated load¹², but no source converts revolutions to hours or kilometres for a duty cycle, so you cannot compute a service life in years (supplier-reported).

Sealing is also the primary life-extending option because it excludes contamination, and thread guards and track brush wipers keep debris from blocking wheels^16,19. If your site cannot hold a grease schedule, specify sealed and stop pretending the zerk will be used.

Confirm washdown and corrosion before the quote

Washdown and chemical exposure attack the bearing and the raceway directly, so a caster chosen on load and push force alone fails in service. The fix is a construction change — stainless, sealed, NSF — not a maintenance change, and it is far cheaper to add before the quote than after.

Sealing the bearing to exclude contamination is the primary option that extends caster life, and thread guards and track brush wipers keep debris from blocking wheels^16,19. One supplier states its 2309 stainless caster is NSF certified and resistant to moisture, corrosion and most chemicals²⁶.

The sources carry no temperature range or chemical-compatibility data per bearing type, so those answers must come from the supplier.

• ✓List every fluid the caster will meet: washdown water, oils, solvents, hydraulic or cutting fluid.
• ✓Ask whether the bearing and raceway are stainless or sealed, not just the bracket.
• ✓Ask for NSF certification if the caster enters a food, pharma or hospital area.
• ✓Ask for the operating temperature range of the specific bearing, not the caster family.
• ✓Ask whether thread guards or track brush wipers are needed for the debris on your floor.
• ✓Confirm the corrosion specification in writing before the bearing type is frozen.

Size the rating with the right safety factor

A published load rating is a test-derived number carrying assumptions. One source states dynamic load capacity is established by validation testing using speed, maximum obstacle height, floor condition, rest/run time and wheel size, and is calculated on one less wheel⁸.

Another states a trustworthy supplier publishes load capacity alongside wheel diameter, bearing type, bracket material, operating temperature and speed limitations, and explains the assumptions behind the rating⁵. Read a rating without those and you are comparing incomparable numbers.

Safety factors run from 1.0-1.5 for manual indoor use with obstacles under 5% of wheel diameter, up to 2.0-3.0 for motorized outdoor use over 5% obstacles²⁷. Those factors do not account for tread wear²⁷.

So add a second margin: one source reports rolling resistance rises 10-20% as wheels wear flat spots or accumulate debris, and bearing drag rises as lubrication degrades, so a 25% resistance margin should be planned at end-of-life²¹.

For a manual indoor cart with obstacles under 5% of wheel diameter, the safety factor is 1.0-1.5 and the wear margin is 25% on top^21,27 (supplier-reported). The sources do not state which component becomes limiting, so ask the supplier to name it.

Put the safety factor and the 25% margin on the RFQ line, not in your head (supplier-reported).

Compare cost per operating hour, not unit price

A longer-life bearing avoids the maintenance labour and downtime that dominate cost. One supplier estimates a higher-priced option at 5-10% more purchase cost returns 10-20% lower total cost of ownership over three years, with maintenance intervals 20-30% longer²².

Treat that as a supplier estimate, not a measured result. The mechanism is the interval: if the greased bearing needs attention every 30-90 days and the sealed unit needs none, the labour and downtime line moves even when the parts line does not¹⁴ (supplier-reported).

The same source states the relevant comparison is cost per operating hour, not unit price, because longer intervals and lower failure rates affect fleet availability²². For a fleet, one hour of downtime on a lift AGV can delay an entire order fulfilment zone²².

The sources carry no unit price, price delta or volume break point between bearing types, so you cannot compute the crossover yourself — you can only ask for both prices at your quantity and run the interval difference.

If your wheels are replaced more than once a year at scale, ask the supplier for the three-year cost per operating hour for each bearing option.

Ask the price, MOQ and lead-time questions

Price, volume breaks, MOQ and lead time do not follow from the load rating, so a bearing type chosen on performance alone can be unavailable or uneconomic at the quantity you actually order. The sources carry no price, MOQ or lead-time data per bearing type, so these answers must come from the supplier.

The only lead time found is 7-10 business days for one roller bearing part, and it may be affected by quantity ordered²⁵ (supplier-reported).

Without these answers you select on capacity and discover at order stage that the chosen caster has a minimum order far above the cart build or a lead time that stops the project.

• ✓Ask for unit price at the quantity you actually need, not the catalogue quantity.
• ✓Ask for volume price breaks at the next order size up.
• ✓Ask for the minimum order quantity for the specific caster and wheel configuration.
• ✓Ask for lead time on the required quantity, and whether it changes with the wheel material.
• ✓Ask whether the quoted rating is per caster or for a complete set.
• ✓Ask for the rating basis in writing alongside the price, so suppliers can be compared.

Check mounting and noise before blaming the bearing

No source attributes noise or vibration to bearing type. Noise tracks wheel tread material and hardness: softer materials are quieter and transmit less vibration, while harder materials roll more easily but are noisier^4,11.

Mounting integrity is the other lever. One source states a threaded stem that is not fully seated, or whose threads wear, lets the caster loosen and introduces vibration and misalignment, which affects rolling efficiency and steering feel¹⁸.

A top plate distributes load across a larger area of the frame, reducing localized stress and helping maintain alignment under side loads and frequent starting and stopping¹⁸. So a bearing upgrade bought to fix a noise complaint is aimed at the wrong component.

The sources give no sound pressure level or subjective noise rating for any bearing type, so you cannot specify a decibel target from this evidence. If noise is the driver, change the tread durometer or the mounting first, and ask the supplier which of the two they are quoting.

For a hospital or office cart, that is the line of the quote to attack first.

Where the sources disagree before you quote

The same upgrade is quoted as three different push-force reductions, and the heavy-load rule is contradicted by a general supplier claim. A buyer who takes one figure as fact can specify the wrong bearing or lose the negotiation.

Use the table to decide which figure to verify with the supplier, and ask for the test basis behind whichever number they quote (supplier-reported).

What the sources do not establish

• No absolute push force, coefficient of rolling resistance or drawbar pull figure per bearing type at a stated load and speed
• No rated load capacity table per bearing type and wheel diameter, and no statement of which component becomes limiting
• No service life in hours, kilometres or years per bearing type under a stated duty cycle
• No sound pressure level or noise rating for any bearing type at comparable speed and load
• No temperature range or chemical-compatibility data per bearing type
• No unit price, price delta, volume break point or MOQ per bearing type; only one lead time found, 7-10 business days for one roller bearing part