Lock four properties into the caster spec before you send the RFQ: wheel material and tread hardness, bearing family, dynamic load rating with its safety factor, and the rolling-resistance and breakaway figures your drive is sized against.
The sources settle material, bearing, load-factor and standard-scope choices, but they publish no numeric diameter tolerance, durometer, breakaway torque or IP value — so demand those from the supplier as declared values. Everything else on the datasheet is a comparison you can make yourself.
Which wheel material should you specify?
Material sets the baseline resistance before any bearing or geometry decision, because a softer tread compresses under each wheel load and the energy lost in that deformation is what raises both the force to start the wheel and the force to keep it rolling2,19.
The trade is per aisle surface, not per fleet: a nylon wheel on a hard smooth floor resists compression and gives the lowest rolling resistance, while the same size polyurethane wheel compresses under load, rolls harder but absorbs shock and protects the floor7,19.
For a 200-piece fleet crossing both a polished concrete aisle and a tiled cell, that means one material choice cannot serve both surfaces well.
| Wheel material | Rolling resistance and compression under load | Choose it when |
|---|---|---|
| Nylon | Harder, resists compression; lowest rolling resistance on hard floors | Hard, smooth floors and high load capacity |
| Polyurethane | Compresses under load; higher rolling resistance, absorbs shock | Floor protection and quiet running matter more |
| Elastic rubber | Stronger shock absorption; lower load for the same wheel size | Shock and vibration dominate the duty |
Which bearing family suits a powered AGV?
The bearing is where internal friction is generated before the wheel ever turns, so a plain bore converts motor torque into heat and axle wear on every cycle, while a precision sealed ball bearing keeps that loss low and repeatable over long runs11,26.
Bearing families rank by rolling resistance from plain (high) through Delrin (medium) and annular ball (low) to precision sealed ball (lowest), and precision sealed bearings are the family recommended for long runs, powered equipment and AGVs11,26.
Plain bearings are for light duty, rarely moved loads where ease of starting and rolling is not important, and they carry no recommended maintenance program11.
One supplier's field note makes the cost of getting this wrong concrete: a 1,000 lb die cart on plain bearings pushed 200 feet per cycle wears the axle out in under 90 days, and the wheel is not the problem26.
That is a bearing failure, not a wheel failure, and it repeats on every unit in the fleet. Delrin is the washdown exception — suited to brine, steam cleaning or excessive water, with light oil or grease lubrication recommended11.
Ask your supplier which family the quoted wheel carries before you compare prices, because the bearing is invisible in a photograph.
What load rating and safety factor go in the RFQ?
Dynamic load capacity is not a fixed property of the wheel.
It is established by validation testing using speed, maximum obstacle height, floor surface condition, rest/run time and wheel size, and it is calculated on one less wheel, so a rating copied from a static table understates the load each wheel actually carries in motion24.
The safety factor you apply depends on where the platform runs: one source gives 1.4–2.0 for indoor power-driven transport with obstacles smaller than 5% of wheel diameter, and 2.0–3.0 for outdoor power-driven transport23. The same wheel therefore needs two different ratings for the same platform, depending on whether it stays indoors.
A second source frames the allowance differently, as a 25–30% margin added to calculated capacity per caster4, and a third gives a general range of 1.0–3.0 with 1.33 as a common allowance23 (supplier-reported figure).
Those are not the same instruction, and the disagreement is the point: the factor is condition-dependent, so pick the one that matches your duty and write it into the RFQ. For an indoor AGV, the 1.4–2.0 band is the one to defend (supplier-reported figure).
Ask the supplier whether the quoted figure is static or dynamic, and at what speed it applies.
Does swivel geometry change path tracking?
Swivel offset, race construction, wheel diameter and load all influence steering effort and shimmy, so an offset chosen for a manual cart can produce wander when the same caster is pushed at AGV speed down a long aisle29. Layout compounds it.
Four-swivel arrangements give a small turning envelope, which suits work cells and crowded rooms, but they can wander in long aisles; a two-rigid plus two-swivel arrangement stabilises a long cart for straight travel29.
For a four-sided bracket running a long straight aisle between two work cells, that is a real conflict: the layout that turns tightly in the cell is the layout that tracks worst on the run between them.
The sources do not give numeric swivel offset, trail or swivel radius values, so you cannot calculate a minimum turning radius from them — ask the supplier for the offset and the race construction on the exact drawing, and test the layout on your own aisle before committing the fleet.
Directional locks are one way to make a swivel caster behave like a rigid caster when engaged, which lets the same platform manoeuvre in a loading area and track through a corridor29. Confirm the mechanism on the drawing, because terminology varies by manufacturer.
Will the caster actually bolt onto the platform?
A bolt pattern that differs by a few millimetres cannot be drilled out or shimmed on a production AGV frame, so a dimensional mismatch discovered at installation stops the whole build rather than degrading performance30.
Wheel "size" refers to diameter, not overall assembly height, and stem types vary between grip ring, threaded and friction fit — three ways to order the wrong part while believing the size matched30. Measure the platform, not the catalogue, and put the numbers on the RFQ drawing (supplier-reported).
- ✓Measure the existing mounting plate length, width and bolt-hole spacing, and record the pattern, not just the plate size
- ✓Confirm the stem type — grip ring, threaded or friction fit — against the platform's socket
- ✓Check overall assembly height against the deck height you need, separately from wheel diameter
- ✓Ask the supplier to confirm the bolt pattern on the quoted part, since patterns can differ by 5 mm between otherwise equivalent casters
- ✓Verify the load rating applies to the complete caster, not the wheel alone
What diameter tolerance must the supplier declare?
Wheel diameter sets the push force required to roll and the distance travelled per motor revolution, so a lot-to-lot diameter spread or an out-of-round wheel changes both the rolling resistance and the odometry of every unit that receives it18,20.
Larger diameters and harder wheel materials on smooth floors cut push force and support ergonomic push limits, which is why larger low-resistance wheels are preferred for manual movement18,20.
The same diameter logic makes a consistent nominal diameter critical on a powered AGV: if one unit's wheels arrive 1 mm larger than another's, the two platforms travel different distances per revolution and the fleet no longer tracks identically.
The sources give no numeric outside-diameter tolerance, no permissible lot-to-lot variation and no roundness or runout limit, so you cannot read a value off a catalogue. Demand it instead: ask the supplier to declare the outside-diameter tolerance, the roundness limit and the lot-to-lot variation for the quoted wheel, and to state the measurement method.
A trustworthy supplier publishes load capacity alongside wheel diameter, bearing type, bracket material, operating temperature and speed limitations, and explains the assumptions behind the rating25 — a supplier who cannot state a diameter tolerance is not publishing at that level. Put the declared tolerance on the spec sheet before you order.
What rolling and breakaway figures size the drive?
Breakaway force is a multiple of steady-state rolling force, so a drive sized only on the rolling coefficient will stall or draw a current spike every time the AGV starts from rest, and the battery runtime claim collapses20.
One source expresses rolling resistance as a coefficient of 0.01–0.08 lb of push per lb of load at steady speed, with breakaway force running 1.5 to 3 times higher than steady-state rolling20.
That multiplier is the number to design against: a loaded caster that rolls at the low end of the coefficient still demands up to three times that force to break away from rest, and the drive must supply it on every start.
The sources give no rolling resistance as a function of speed, no floor-specific coefficient and no numeric breakaway torque value, so you cannot compute the start-up figure from the material alone. Ask the supplier for the breakaway force or torque at your wheel load, and for the rolling coefficient at your speed and floor.
Write both into the RFQ as declared values, and size the motor and battery against the breakaway figure rather than the rolling one. If the supplier quotes only a rolling coefficient, treat the start-up case as unverified.
What environmental and certification scope applies?
Chemical resistance depends on concentration, contact time, temperature and humidity rather than the chemical name, so a material chosen from a compatibility chart alone can fail in service7.
A stainless bracket does not solve it either: a stainless mounting plate cannot protect a standard bearing from repeated washdown, so bearings, axle hardware and brakes must be verified separately7.
Certification scope is the other half — a capacity number without a test basis is hard to compare, and the standard's speed limit may exclude your platform10.
- ✓State the actual chemical, its concentration, contact time, temperature and humidity — not just the chemical family
- ✓Verify bearings, axle hardware and brakes for washdown, not only the bracket material
- ✓Confirm whether the quoted rating is static or dynamic, and at what speed it applies
- ✓Check whether the complete caster or only the wheel received the rating
- ✓Confirm the application speed falls within BS EN 12532's scope of up to 1.1 m/s, approximately 4 km/h
- ✓Ask whether temperature reduces the quoted capacity
Can you trust the supplier's published rating?
A rating without its assumptions cannot be compared against another supplier's rating, so the buyer ends up comparing numbers produced under different speed, temperature and duty assumptions and picks the wrong caster25.
Reliable manufacturers avoid presenting one maximum number as suitable for every situation and instead show how speed, environment and wheel configuration change performance25. The checks below turn a marketing number into a traceable one.
- ✓Ask which standard supports the rating, and whether the value is static or dynamic
- ✓Ask at what speed the rating applies, and which wheel and bearing configuration was tested
- ✓Confirm the supplier publishes load capacity alongside wheel diameter, bearing type, bracket material, operating temperature and speed limitations
- ✓Ask for the assumptions behind the rating — speed, environment and wheel configuration
- ✓Check whether the rating covers the complete caster or only the wheel
- ✓Confirm the quoted capacity applies to the caster, not to a multi-pack set
Where do the sources disagree?
Two sources stating different figures for the same property means you cannot adopt either as a specification value without checking the application conditions behind each, and the disagreement itself is the signal that the figure is condition-dependent4,7,10,19,23,24.
The rows below keep both sides visible so you can see which figure to confirm rather than assume (supplier-reported).
| Disputed item with unit | One source reports | Another reports | What the buyer should do |
|---|---|---|---|
| Rolling resistance by wheel material | Nylon gives less rolling resistance, especially on hard floors | Polyurethane compresses under load, increasing rolling resistance | Confirm the coefficient on your actual floor and load |
| Load-capacity safety factor | Add 25–30% to calculated capacity per caster | 1.4–2.0 indoors, 2.0–3.0 outdoors for power-driven transport | Pick the factor matching your duty and write it in |
| Dynamic load calculation basis | Calculated on one less wheel | Total load × safety factor ÷ supporting casters | Ask the supplier which basis the quoted rating uses |
| Standard scope for industrial castors | BS EN 12532 covers castors up to 1.1 m/s | No source addresses ISO 22883, ANSI/ICWM or EN 12530 | Confirm the standard on the supplier's test report |
What the sources do not establish
- No numeric wheel outside-diameter tolerance, lot-to-lot variation or roundness/runout limit for AGV/AMR casters
- No rolling resistance as a function of speed or floor type, and no numeric breakaway/start-up torque value
- No durometer (Shore A/D) values for polyurethane, rubber or nylon caster treads
- No numeric dynamic load rating at a specific AGV/AMR speed, and no shock-load allowance value
- No numeric swivel offset, trail or swivel radius values, and no minimum turning radius formula
- No dimensional tolerance values for the mounting interface or overall height
- No IP ratings, operating-temperature ranges, or ISO 22883, ANSI/ICWM or EN 12530 coverage
- No maintenance interval values for bearings beyond qualitative statements
Frequently asked questions
What wheel properties matter when selecting casters for AGV and AMR platforms?
Lock four properties into the caster spec before you send the RFQ: wheel material and tread hardness, bearing family, dynamic load rating with its safety factor, and the rolling-resistance and breakaway figures your drive is sized against. The sources settle material, bearing, load-factor and standard-scope choices, but they publish no numeric diameter tolerance, durometer, breakaway torque or IP value — so demand those from the supplier as declared values.
What is not established about Selecting casters for AGV and AMR platforms?
No numeric wheel outside-diameter tolerance, lot-to-lot variation or roundness/runout limit for AGV/AMR casters. Where this matters to your order, ask the supplier for the specific test or datasheet value rather than accepting a general claim.
What is not established about Selecting casters for AGV and AMR platforms?
No rolling resistance as a function of speed or floor type, and no numeric breakaway/start-up torque value. Where this matters to your order, ask the supplier for the specific test or datasheet value rather than accepting a general claim.
Sources · 15
- 2grainger.comIndustry publication2023-12
- 4thecasterguy.comIndustry publication2024-07
- 7conveyorparts.euIndustry publication2026-09
- 10conveyorparts.euIndustry publication2026-09
- 11thecasterguy.comIndustry publication2023-05
- 15holkie.comIndustry peer technical page2026-09
- 18thecasterguy.comIndustry publication2024-07
- 19monroeengineering.comIndustry publication2026-09
- 20casterhq.comUnclassified source2026-08
- 23wheelswaycaster.comIndustry peer technical page
- 24algood-casters.comManufacturer technical documentation2026-01
- 25conveyorparts.euIndustry publication2026-09
- 26casterhq.comUnclassified source2026-09
- 29de.ylcaster.comUnclassified source
- 30infordcaster.comUnclassified source2026-05
Technical references cited for verifiability — not supplier recommendations.