Why Caster Wheels Flat Spot, and How to Choose Material, Load Rating and Replacement Timing

Flat spots are compression set, not wear

A flat spot is not abrasion. It is the contact patch taking a permanent set because the wheel stood still under load.

Flat spots form most often on polyurethane and rubber wheels that sat parked under heavy load for long stretches, and you feel them as a thumping vibration when the equipment rolls slowly 3 , 14 .

Rubber is the compound with a stated threshold: under heavy static loads exceeding 50% of rated capacity, rubber wheels develop flat spots per the compression-set behaviour documented in ASTM D395 1 .

That is why sizing a wheel against the static weight it will hold, not just the weight it will roll, is the first check you make. The corrective options reported are a harder compound, a larger diameter, or a change to the parking habit 3 , 14 .

If your carts sit loaded overnight or across a weekend, treat the parked condition as the design case. The sources do not publish compression-set or rebound-resilience values per material, so you cannot rank compounds numerically from this evidence — you can only rule out the soft ones and size the load.

Choosing the wheel material for the sustained load

Polyurethane resists compression set better than rubber and carries heavier loads, while soft rubber deforms under load and absorbs energy into the contact patch 1 , 5 . That is the mechanism behind the flat-spot pattern: the softer the tread, the more of the parked load goes into permanent deformation rather than elastic recovery.

Temperature is the second rule-out. Standard polyurethane is the lower-temperature material, and the wheel material — not the bracket — is the limiting factor in high-temperature washdown 4 .

If your duty cycle includes hot washdown or hot product contact, check the material's temperature limit before you compare anything else.

| Material | Maximum operating temperature | |---|---| | Standard polyurethane | ~160F 4 | | Polypropylene | ~200F 4 | | Nylon | ~250F 4 | For loads over 500 lbs per caster, polyurethane or harder materials such as nylon or phenolic are recommended 1 .

Soft rubber is the material to rule out where the wheel stands loaded: its rolling coefficient of 0.06-0.08 at 70A is more than double the 0.03 of 95A polyurethane, because it deforms under load and absorbs energy into the contact patch 5 .

Ask your supplier for the compound's compression-set behaviour under your parked load, since the sources give no per-material test values.

Sizing the load rating before you quote

Undersized ratings are the failure mode that produces flat spots, cracks and broken swivel bearings, and the usual cause is matching the rating to the static cart weight instead of the dynamic load 10 .

The method is to divide the loaded weight by the number of wheels minus one — or by three — and then apply a safety factor 2 , 10 , 19 . Capacity ratings describe intermittent service.

For 24/7 operation, one source reports derating by 25-33% because the rating assumes intermittent duty 30 . Treat that as a source-specific figure to verify against your duty cycle rather than a fixed rule.

Safety factors reported across sources range from 20% to 1.5x depending on application, floor condition and movement frequency 2 , 10 , 19 , 21 , 35 , 38 . The sources do not agree on a single value, so pick one that matches your floor and duty cycle and write it into the RFQ.

Deciding whether a flat-spotted wheel can be recovered or must be replaced

Mild flat spots can recover by rolling; severe ones cannot. Measure runout before you decide, and remember that any visible flat section is structural failure regardless of the measurement 6 .

• ✓Measure runout: 0.005-0.015 in is mild — roll the cart 2-4 hours on a smooth floor, most recover within 24 hours, re-inspect at 72 hours.
• ✓0.015-0.030 in is moderate — it may recover partially but will flat-spot again faster; plan replacement within 30 days and rotate to a low-hour cart.
• ✓Over 0.030 in is severe — permanent plastic deformation; replace immediately to avoid bearing damage and vibration.
• ✓Any visible tread deformation is structural failure — replace regardless of measured runout.
• ✓After replacement, inspect the bearing for rough spin or debris; flat-spotted wheels vibrate and damage bearings.
• ✓Set the inspection interval to the duty cycle: monthly in pharma, every 3-6 months in high-use industrial, annual in light commercial.
• ✓Replace on flat spots, cracks, tread separation, wobble, or loss of rated load capacity.

Diameter, tread and bearing choice for the duty cycle

Rolling resistance is what decides push force, and it is driven mostly by tread material and diameter rather than bearing type. Bearing choice matters less.

Precision ball bearings have 20-30% less rolling resistance than roller bearings, while tapered roller bearings handle heavier load but add 10-15% resistance versus ball bearings 5 . If your operators push long distances, that difference is worth checking; if the cart mostly stands loaded, it is not the deciding factor.

Match the wheel to the route as well as the load. A wheel sized correctly for the load can still be too small for the floor it crosses, so walk the full route with the actual load before you fix the diameter 10 .

Standards and test data to require in the RFQ

Several standards are named for caster and wheel testing, including ICWM/ANSI MH31.1, ISO 22883, DIN EN 12532, ABMA 9, ASTM B117 and ASTM D395 1 , 15 , 18 , 34 . Naming them in the RFQ is useful because it tells the supplier which test framework you expect the data to come from.

What the sources do not provide is the actual required load and life test data or the pass criteria behind those standards 15 , 18 , 34 , 36 . So the RFQ should ask for the test report itself, not just the standard number: which test was run, at what load, for how many cycles, and what the acceptance criterion was.

A quote that names a standard without the report behind it does not let you compare suppliers.

Normalising quotes before you compare price

Caster quotes are only comparable when they describe the same thing. Ask every supplier to quote against the same load, wheel diameter, brake mode, mounting dimensions, tread requirement, quantity and documentation request — a quote with a central lock and one with a simple full brake are not the same product 8 .

Then clarify what sits outside the unit price: sample cost, tooling or customisation charges, packaging, test reports and lead time, plus the order quantity the quoted unit price assumes 8 . A lower unit price can disappear once frame changes, push resistance or floor marking are added to the comparison 8 .

For OEM projects, sample testing and drawing confirmation are often worth more than a small unit-price difference 8 . The sources do not publish MOQ or custom-mould tooling figures, so ask for them explicitly rather than assuming a range.

Where the sources disagree

Two decisions in this article rest on figures the sources do not agree on. Keep both positions visible and verify against your own duty cycle rather than picking a number because it appears more often.

What the sources do not establish

• No source provides compression-set or rebound-resilience test values per wheel material, so a quantitative material ranking cannot be established.
• No source provides glass-transition or softening-point data for wheel materials.
• No source gives quantified travel distance, speed, or duty-cycle thresholds for flat-spot formation.
• No source provides the actual required load/life test data or pass criteria for the named standards (ICWM/ANSI MH31.1, ISO 22883, DIN EN 12532, ABMA 9, ASTM B117).
• No source gives MOQ or custom-mould tooling cost figures.
• No source provides a quantified relationship between wheel diameter or tread width and flat-spot resistance.
• No source provides UV exposure degradation data for wheel materials.
• No source provides a load-capacity-loss percentage that triggers replacement.
• No single mandatory safety factor is established; the sources report different values for different conditions.
• The 0.005-0.030 inch runout thresholds come from a single industry-tier source and are not corroborated by a governing standard.
• No single inspection interval is universally required; the sources give different intervals for different duty cycles and environments.