Caster Wheel Material by Floor, Load and Environment: What the Evidence Settles

Start with the worst-case load per caster, not the average

Sizing starts from total load divided by three, because one caster can be off the ground on an uneven floor, and a safety margin is added on top of that figure 1 , 4 , 33 . The margin is a recommendation, not a governing standard: one source sets a 20% minimum 1 and another recommends 25% 4 .

Shock loading then multiplies the effective load again, with dynamic multipliers of 1.5x–2x reported for shock conditions 33 . That sequence matters because a wheel that passes on paper can still be overloaded in service.

The material ranking is consistent across sources — steel and cast iron carry the most, then nylon, then PU, then rubber 5 , 11 , 18 , 24 .

Rubber is the binding constraint: most rubber wheels are rated between 200 and 500 lb per caster, and above 500 lb per caster the recommendation shifts to PU or harder materials such as nylon or phenolic 2 , 5 . PU is cited at up to 2,000+ lb per caster 2 , 5 .

So the first cut is arithmetic, not preference. If the worst-case per-caster figure after the margin and the shock multiplier lands above the rubber band, rubber is out before floor type is even discussed.

Eliminate materials the floor itself rules out

Harder wheels concentrate contact pressure into a smaller patch, and that pressure is what marks, scratches or indents a soft floor; softer wheels spread the same load over more contact area 8 , 17 , 18 . The working principle stated across sources is that the harder the floor surface, the softer the caster wheel should be 17 , 18 , 23 .

Applied to specific floors, the recommendations converge on PU and rubber/TPR for delicate surfaces. Soft PU or TPR/TPE is named for hardwood, rubber or TPR for tile, PU for vinyl, and soft PU or neoprene for epoxy 17 .

PU is described as non-marking on concrete, epoxy, tile and most commercial flooring, while rubber is gentler still 2 .

Nylon and steel are the materials to keep off finished floors: nylon marks soft floors and is suited to concrete and epoxy, and steel or cast iron damages most floors and is reserved for industrial settings where floor condition is not a concern 8 , 11 , 19 , 26 .

One practical consequence: a floor-damaging wheel should never reach the shortlist, because the cost of the floor repair is not carried in the wheel price.

If the floor is finished and the load is high, the conflict between floor protection and capacity is real, and the resolution is usually a larger-diameter softer wheel rather than a harder one.

Compare the surviving materials on load, floor behaviour, noise, environment and cost direction

Once the load cut and the floor cut are applied, the remaining materials are compared on the same axes. Cost direction is a direction only — the sources rank materials relative to each other and publish no prices, so treat the cost column as an ordering, not a budget.

Check the environment before locking the material

Temperature is the first hard limit. PU is rated from -40°F to 200°F and rubber from -20°F to 180°F, so rubber gives up roughly 20°F of headroom at the top end 5 .

Above those ceilings the recommendation moves to phenolic or steel, which are named for high-heat service 4 , 5 , 19 . Chemical exposure separates the two soft materials sharply: PU is rated high for resistance to oils and solvents, while rubber is rated low and degrades with chemicals 5 .

That single difference often decides between them in a machine shop or a washdown area, where the floor protection argument would otherwise favour rubber. Wet, corrosive and cleanroom environments push toward stainless steel or nylon for corrosion resistance 13 , 19 , 22 .

Outdoor and rough-surface use adds its own requirements — weather-resistant materials, anti-rust treatment and a larger wheel diameter 23 . Note that the sources give temperature ranges only for PU and rubber; no range is established here for nylon, steel or cast iron, so those limits have to come from the supplier.

Check flat-spotting risk against the duty cycle

A wheel that is correctly rated can still fail if it sits loaded. Rubber develops flat spots under sustained static load above 50% of its rated capacity, through compression set 2 , 5 .

PU resists flat spotting at 60% or more of rated load 5 . The gap between those two thresholds is the reason a cart that parks loaded overnight behaves differently on rubber than on PU, even when both are within their load ratings.

Wear rate compounds the same decision. PU is reported to last 3–5x longer than rubber 2 , 5 , which is the basis for the cost-direction claim that PU costs more upfront but less over time 5 .

For a fleet that parks loaded, the practical check is the ratio of static load to rated capacity, not the peak load alone. If that ratio sits above half the rating on rubber, either move to PU or plan an inspection interval that catches compression set before it becomes a ride-quality complaint.

Adjust diameter and bearing for the actual floor surface

Material alone does not set how hard a cart is to push. The same caster measures different rolling resistance on different floors, and the multipliers are large: polished epoxy-coated concrete is the baseline at x1.0, unsealed concrete x1.2, expansion joints x1.5–2.0, rough asphalt x1.8, and industrial carpet or anti-fatigue mat x2.5–3.5 3 .

The mechanism is that soft or irregular floors deform under the contact patch or force the wheel to climb micro-discontinuities, sinking energy into the floor instead of into rotation 3 . Load adds to this directly, with a linear relationship between load per caster and rolling resistance 9 .

Diameter is the lever. Larger wheels distribute load over a larger area and roll more easily over obstacles 14 , 24 , which is why expansion joints and carpeted floors call for a larger diameter rather than a different material 3 .

Bearing choice can matter as much as the wheel material. Bearing load limits range from 900 lb for Delrin to 3,500 lb for roller bearings, with plain bearings up to 1,200 lb 7 .

A material that passes every other check can still be the wrong specification if the bearing is undersized for the load.

Verify certification and cleanability before the material is locked in

In food, medical and cleanroom service, certification and cleanability can override the material comparison. NSF certification is available on specific stainless and PU products, and it is listed alongside CA Prop 65 as a certification consideration for caster selection 12 , 31 , 32 .

• ✓Confirm whether the application requires NSF certification or another standard, and ask for the certificate rather than a catalogue claim.
• ✓For food processing or medical use, ask for NSF-certified or stainless products and easy-clean materials.
• ✓Check whether the wheel is non-marking and compatible with stainless casters if hygiene and floor appearance both matter.
• ✓Ask for the supplier's stated temperature range for the specific wheel, since ranges are published for PU and rubber but not for nylon, steel or cast iron.
• ✓Ask how the wheel behaves under sustained static load, and whether the supplier's rating accounts for compression set.

Where the sources disagree, and what to ask

Two disagreements survive across the sources. Both concern PU, and neither is resolved by the evidence in this set, so they belong in the supplier conversation rather than in the selection logic.

What the sources do not establish

• No load-vs-diameter table for each material; diameter-specific load ratings are not established.
• No quantitative push or pull force values per material, load and floor surface.
• No expected service life in hours or miles for any material.
• No specific prices, MOQs, tooling costs or lead times per material.
• No FDA or USDA certification references; only NSF is mentioned.
• No temperature ranges for nylon, steel or cast iron.
• No chemical resistance data against specific chemicals for each material.
• No quantitative rebound or noise (dB) values.
• No consolidated selection chart combining all five materials across all floor types, load ranges and environments.
• The 20% and 25% safety margins are source-reported recommendations, not governing standards; no mandatory industry figure is established.