Heavy equipment dollies need three separate checks: the load carried by each support, the additional demand from impacts, and the concentrated loads applied to the floor. Adding the capacities printed on four dollies does not establish that they can move a machine of that total weight.
Start with the machine’s loaded weight, center of gravity, support locations, and travel route. Then check the complete dolly or caster assembly under its intended operating conditions. A wheel rating alone does not establish the capacity of its axle, swivel assembly, mounting plate, dolly frame, or attachment to the equipment.
The governing value is the highest credible load on an individual support—not the average load across the set.
Establish What Each Capacity Rating Applies To
A machinery skate, a recirculating-chain roller, and a caster-mounted platform dolly distribute loads differently. Identify the equipment type before dividing the load.
| Published rating | What it describes | What still needs confirmation |
|---|---|---|
| Individual wheel capacity | The specified wheel and bearing configuration | Axle, caster bracket, mounting, and dolly-frame capacity |
| Complete caster capacity | The specified wheel-and-bracket assembly | Attachment strength, dolly structure, and application conditions |
| Individual dolly or skate capacity | One complete support unit | Load distribution among units and permitted loading of its top surface |
| Matched kit or system capacity | The manufacturer’s specified combination | Required arrangement, accessories, support conditions, and operating restrictions |
Use the exact model’s documentation. Do not multiply a kit rating by the number of pieces, or assume that a dolly’s load divides equally among its internal wheels or rollers.
Record both the standing and moving conditions:
- Machine weight, including fluids, tooling, fixtures, and anything remaining on the machine during transport.
- Moving-platform weight and any temporary beams or adapters carried by the wheels.
- Support locations and the loaded center of gravity.
- Manual movement or powered towing, including intended speed.
- Floor joints, thresholds, slopes, covers, ramps, and changes in level.
- Expected stationary dwell time on the dollies.
In the calculations below, weight and support reactions are forces expressed in lbf. A manufacturer may label load capacity “lb”; confirm that this represents supported weight rather than a shipping mass. A mass stated in kilograms must be converted to force before combining it with force-based equations.
Calculate Support Loads Before Selecting a Dolly Set
For a preliminary vertical-load model, the support reactions must balance the total weight:
[
\sum R_i=W
]
where (R_i) is the vertical reaction at support (i), and (W) is the total supported weight, both in lbf. Moment equilibrium must also be satisfied; the reactions depend on where the center of gravity lies relative to the supports.
Dividing (W) by the number of supports is valid only when the arrangement and loading justify equal sharing. Blickle specifically notes that individual loads can vary when four or more wheels or casters are used. Its application allowances also depend on operating conditions rather than a single universal factor.
Example: An Off-Center Machine on Four Supports
Assume a hypothetical machine and moving platform weigh 12,000 lbf together. Two front supports and two rear supports are separated by distance (L). The loaded center of gravity is (0.60L) forward of the rear support line.
On a level surface, with no horizontal force or acceleration:
[
R_{\text{front pair}}=W\frac{x}{L}
=12{,}000(0.60)=7{,}200\text{ lbf}
]
[
R_{\text{rear pair}}=W-R_{\text{front pair}}
=4{,}800\text{ lbf}
]
Here, (x) is the horizontal distance from the rear support line to the center of gravity, measured in the same units as (L).
If the load is laterally centered and the support arrangement provides equal sharing within each pair, each front support carries 3,600 lbf and each rear support carries 2,400 lbf. The simple four-way average of 3,000 lbf understates each front support’s load by 600 lbf.
The pair reactions follow from longitudinal moment balance. The individual reactions still depend on lateral balance, frame stiffness, support heights, and contact with the floor.
Account for Uneven Floors Without Assuming a Universal Rule
Four rigid supports may not remain equally loaded across a twisted floor or a joint. One support can unload while others take more weight. Replacing the support count with “number of wheels minus one” is not a complete load-distribution analysis: it does not locate the center of gravity or identify which support takes the increased reaction.
A three-point arrangement avoids four-point rocking, but it does not automatically divide weight into thirds. Its reactions depend on the center of gravity’s position within the support triangle.
Where individual loads remain uncertain, obtain a support-load analysis or use appropriately rated load cells or weighing equipment under a controlled procedure. Measurements on one level location do not establish the maximum reactions along an uneven route. Do not reposition or instrument a heavily loaded dolly without an approved lifting and support method.
Check Standing and Rolling Capacity Separately
A static capacity applies to the manufacturer’s specified stationary conditions. A dynamic capacity applies to specified rolling-test conditions. Neither is a substitute for the other.
For example, Blickle describes dynamic capacity in terms of rolling-bench testing and identifies tread flattening over time as one consideration in its static-load testing. These descriptions explain the distinction; they do not establish the ratings of an unrelated dolly or wheel.
For an individual caster:
[
U_s=100\frac{R_s}{C_s}
\qquad
U_d=100\frac{R_d}{C_d}
]
where:
- (U_s) and (U_d) are static and dynamic utilization, in percent.
- (R_s) and (R_d) are the applicable standing and rolling reactions, in lbf.
- (C_s) and (C_d) are the corresponding manufacturer-published capacities, in lbf.
Using the hypothetical 3,600-lbf front-support reaction, assume a complete caster has a 6,000-lbf static capacity and a 4,500-lbf dynamic capacity. If 3,600 lbf represents both load cases:
[
U_s=60%
\qquad
U_d=80%
]
These results describe how much of each stated rating the assumed reaction uses. They do not demonstrate suitability for impacts, towing, slopes, or a different temperature or duty cycle. An 80% utilization is not a universal acceptance threshold.
The Static vs Dynamic Caster Capacity Calculator provides one caster-load input and separate static and dynamic rating inputs. Use it with verified ratings for the same assembly. If standing and rolling reactions differ, evaluate those cases separately rather than treating one input as both. The tool does not model impact or automatically adjust ratings for speed, temperature, or continuous duty.
If the manufacturer publishes only one capacity, establish what that rating covers. Copying it into both fields does not create a verified second rating.
Treat Set-Down Shock as a Separate Load Case
Lowering equipment onto a dolly, dropping into a floor gap, and striking an upward threshold are different events. Their forces depend on the moving mass, velocity, stiffness, damping, and load sharing.
A vertical free-drop model can illustrate a set-down event. It cannot directly predict the force from a wheel rolling into a raised edge.
A Defined Linear-Spring Model
For a rigid weight falling vertically onto an initially unloaded, linear-elastic support, an idealized peak force is:
[
F_{\text{peak}}
=P\left(1+\sqrt{1+\frac{2h}{\delta_s}}\right)
]
where:
- (P) is the weight participating in the impact, in lbf.
- (h) is the free-fall distance before contact, in inches.
- (\delta_s) is the support’s static deflection under (P), in inches.
- (F_{\text{peak}}) is the estimated peak spring force, in lbf.
The model equates the falling weight’s work to stored elastic energy. Michigan State University’s impact mechanics notes present this relationship with assumptions including negligible support mass, negligible deformation within the impacting mass, and no damping. It is a mechanics model, not a caster-rating test.
Static deflection is not the same as maximum impact travel or travel remaining before bottoming out. Substituting one for another changes the model.
For a hypothetical isolated support carrying (P=500) lbf, with (h=0.25) in and (\delta_s=0.05) in:
[
\frac{F_{\text{peak}}}{P}
=1+\sqrt{1+\frac{2(0.25)}{0.05}}
\approx4.32
]
[
F_{\text{peak}}\approx2{,}160\text{ lbf}
]
The corresponding peak deflection is:
[
\delta_{\text{peak}}
=\delta_s\frac{F_{\text{peak}}}{P}
\approx0.216\text{ in}
]
If the real support bottoms out before that deflection, or its stiffness changes substantially, this model no longer represents its response. The 500-lbf example also assumes that the entire stated weight participates in that support’s impact; it does not establish the effective impacting weight of a connected machine.
At zero drop height, this idealized model gives twice the weight because it represents sudden application to an unloaded spring. A load lowered gradually is a different case.
Use the Shock Calculator Without Treating It as an Impact Rating
The Caster Shock Load Calculator accepts per-caster load, drop height in inches, and an assumed deflection in inches, then reports an equivalent impact load and multiplier. Its accessible page does not expose enough calculation detail to confirm that its deflection input matches (\delta_s) in the model above. The example above therefore defines its own model rather than claiming to reproduce that tool’s output.
Do not select a deflection solely from a tread-material label. Obtain load-deflection information for the specific assembly and load range when the result affects equipment selection.
Likewise, a calculated peak force below a catalog dynamic capacity is not proof of impact suitability. Ask the manufacturer whether the assembly is suitable for the stated set-down event and which published rating or test supports that use. If impact conditions cannot be established, remove the drop or change the transfer method rather than relying on an arbitrary multiplier.
Calculate Wheel Contact Pressure From the Actual Patch
Average contact pressure describes force distributed over the wheel’s floor-contact area:
[
p_{\text{avg}}=\frac{R_w}{A}
]
For an approximately rectangular patch:
[
A=b_c l_c
\qquad
p_{\text{avg}}=\frac{R_w}{b_c l_c}
]
where:
- (R_w) is the individual wheel reaction, in lbf.
- (A) is the contact area, in square inches.
- (b_c) is the actual contact width across the tread, in inches.
- (l_c) is the contact length along the direction of travel, in inches.
- (p_{\text{avg}}) is average pressure, in psi.
Use the footprint under the relevant load. Nominal tread width may exceed the width actually touching the floor, particularly with a crowned tread. A rectangular estimate also does not establish the local peak pressure within the patch.
For the hypothetical 3,600-lbf wheel load, assume an actual rectangular contact patch 3.0 in wide and 0.50 in long:
[
A=3.0(0.50)=1.5\text{ in}^2
]
[
p_{\text{avg}}=\frac{3{,}600}{1.5}
=2{,}400\text{ psi}
]
The equivalent value is 345,600 psf because (1\text{ psi}=144\text{ psf}). This is the same local contact pressure expressed in different units—not a distributed load over one square foot of floor.
The Caster Floor Point Load Calculator uses wheel load in lbf, tread width in inches, and contact length in inches, and reports pressure in psi and psf. Its contact patch is an input, not a prediction of tread deformation. Use a manufacturer-supplied or appropriately measured footprint rather than an unverified default.
For a multiwheel dolly or chain roller, determine the loads and actual contact areas of the wheels or rollers touching the floor. The dolly frame’s outer dimensions are not its ground-contact area.
Separate Surface Damage From Floor Structural Capacity
Wheel contact pressure and floor load capacity answer different questions.
A finish, tile, coating, or resilient flooring product may be limited by indentation, cracking, bond failure, rolling traffic, or stationary dwell. A structural floor must also resist the concentrated wheel forces and their arrangement.
Do not compare wheel pressure in psf directly with a mezzanine’s uniformly distributed load rating. Converting units does not make the loading conditions equivalent. ACI’s published abstract for “Design of Concrete Slabs for Concentrated Loads” distinguishes concentrated-load effects from uniform loading and includes a forklift-loaded slab example; that abstract does not supply a permissible wheel load for a particular building.
For a floor review, provide:
- Maximum wheel or roller reactions and their spacing.
- Contact dimensions under the relevant load.
- Travel path, turning locations, and parking locations.
- Known impacts, floor joints, slab edges, covers, and openings.
- Available floor drawings, construction details, and condition records.
A larger wheel footprint can lower average surface pressure without reducing the total force delivered to a floor beam or support. Similarly, a steel running plate is not automatically an adequate load spreader: its thickness, span, support, edges, and resistance to movement determine how it transfers load.
Where floor construction or concentrated-load capacity is unknown, obtain a floor assessment before moving the equipment.
Resolve Route and Equipment Limits Before the Move
A dolly set can satisfy the preliminary capacity calculations and still be unsuitable for the route.
Confirm that the load-bearing top surfaces fit the machine’s approved support points, that attachments prevent unintended movement, and that the loaded center of gravity remains within the support arrangement throughout turns and transitions. A high vertical rating does not establish lateral restraint or tipping resistance.
Also confirm the intended movement method. Static capacity does not establish towing speed, braking ability, drawbar suitability, or the force needed to start and steer the load.
If a wheel or caster replacement is proposed, check more than diameter and capacity. Installed height, mounting geometry, axle fit, swivel clearance, brake interference, and load sharing must remain compatible with the dolly.
A suitable preliminary selection therefore needs an identifiable load path: from the machine’s support points, through the complete dollies or caster assemblies, into verified wheel-contact areas, and finally into a floor capable of carrying those concentrated loads. Any unresolved link requires specific manufacturer data, measurement, or engineering review before the move proceeds.