Engineering decision guide · Sources checked 2026-11-02
Published by AMRDriveUnit. Page created , last reviewed . Review every six months and when component documentation or standards change. Repository creation date does not establish first public publication.
Custom drivetrain: decisions after the first estimate
Use the tool to frame an AMR/AGV drivetrain RFQ. Use this report to distinguish calculated demands, illustrative assumptions and evidence still needed for component selection.
Load path
Confirm wheel-hub reactions before choosing the reducer
The tool estimates normal load per hub. Overhang, bearing spacing, axial force and preload remain unknown. Calculate individual bearing reactions and obtain the selected bearing’s ratings before accepting a life estimate.
Evidence and limitsComponent selection
Treat target reduction as a starting point
The ratio follows the entered motor rated speed and wheel speed. It is not a stocked gear ratio or a torque-speed approval. Compare available ratios and CAD envelopes; no topology guarantees a chassis height.
Evidence and limitsMaterial evidence
Wheel family alone cannot approve speed or cold operation
Load, diameter, tread thickness, formulation, duty and ambient temperature all require supplier confirmation. The tool uses declared example coefficients and speed screens; it does not predict tread temperature or certify freezer use.
Evidence and limitsVehicle validation
Separate static holding from dynamic stopping
Ramp-hold torque does not determine stopping distance. Use measured worst-case braking and the installed scanner’s response times and supplements for the safety field. No protective field number is issued here.
Evidence and limits1. Reproduce the calculation and inspect its assumptions
Let m be payload plus tare, r the wheel radius, n the driven-wheel count and θ = atan(grade / 100). The steady force is m g (Cr cos θ + sin θ); the peak adds 1.12 m v / acceleration time. Per-wheel torque is force × r / n. This is an uphill, straight-line concept model; it does not cover towed loads or turning.
Electrical power is mechanical force × velocity divided by gear, motor and inverter efficiencies; total propulsion DC current is electrical power / bus voltage. Losses at each stage equal its input power minus output power. A peak at top speed is an envelope estimate, not a controller or stall-current rating. The entered rated motor rpm is treated as the allowed speed ceiling.
What the calculation knows and assumes
All coefficients below are author-selected examples, not measurements or standards requirements.
| Quantity | Model treatment | Decision limit |
|---|---|---|
| Mass | Carried payload + vehicle tare | Trailer/towed mass must use a separate model |
| Grade and acceleration | Simultaneous maximum uphill grade and linear acceleration; rotary factor 1.12 | Excludes turning scrub, yaw, wind and obstacle impact |
| Efficiency | Gear: 0.94 / 0.91 / 0.88 / 0.82 / 0.97 by candidate; motor: 0.92 / 0.88 / 0.83; inverter: 0.96 | Constant efficiencies omit stall, low-speed and temperature behavior |
| Normal load and traction | Equal share among driven wheels; assumed μ × driven normal force | No load transfer, wet-floor measurement or tire-force direction model |
| Illustrative bearing life | C = 28,500 N; P = 1.35 × hub radial load; roller exponent 10/3 | No selected bearing, reaction geometry or service-life prediction |
| Holding brake | 1.5 × grade wheel torque / reduction | Author-selected multiplier; no backdrive losses or compliance approval |
| Inertia and lost motion | Translational mass divided equally per motor; 0.6 m example track width | Excludes rotating inertia; heading illustration applies only to a two-wheel differential vehicle |
| Energy | Cruise current × duty fraction × operating hours | Excludes auxiliaries, idle, acceleration energy, regen and battery derating |
2. Choose the transmission around loads and CAD
Transmission candidates and tradeoffs
Architecture guidance, not product ratings. Efficiency, backlash, shock capacity and envelope require an exact part number.
| Candidate | Reason to consider | Tradeoff / mismatch | Evidence needed |
|---|---|---|---|
| Coaxial planetary | Aligned motor and wheel axes; conventional reduction | Axial package may interfere with the lift or chassis | CAD envelope; efficiency and torque curves |
| Right-angle bevel-planetary | Change motor orientation for packaging | Additional mesh and bearings; cost and service access | Full envelope including wheel; losses at duty point |
| Cycloidal | High reduction and lost-motion requirements | Efficiency, mass and torsional behavior vary by design | Supplier peak, emergency and fatigue ratings |
| Strain-wave | Compact high reduction; precision motion | Flexspline fatigue and compliance need duty review | Allowable load spectrum and thermal behavior |
| Direct-drive hub | Eliminate reduction gears and gear backlash | Motor must supply wheel torque at wheel speed | Low-speed torque/current, cooling and bearing ratings |
A low chassis may benefit from changed motor orientation, but the wheel itself still needs physical clearance. Do not infer a 120 mm total height from a 160 mm wheel or a named topology. Compare real component envelopes and maintenance access. See custom drive assemblies for mounting interfaces.
3. Resolve bearing reactions before fatigue life
For a roller bearing at constant speed, L10h = 10⁶ / (60 × rpm) × (C/P)^(10/3). The tool substitutes C = 28,500 N and P = 1.35 × hub radial load to illustrate sensitivity only. It does not resolve actual bearing reactions or certify a dual-bearing hub. Basic L10 uses 90% reliability and excludes several non-fatigue failure mechanisms. NSK formula; ISO scope.
Hub-bearing review before a life calculation
ISO 281 and the NSK explanation describe rating life. They do not specify which arrangement fits your vehicle.
| Arrangement / input | Potential benefit | Verification needed |
|---|---|---|
| Deep-groove ball bearings | Simple packaging where loads permit | Radial/axial ratings, fits and overhang reactions; use exponent 3 |
| Opposed tapered rollers | Combined-load support and adjustable preload | Each bearing’s radial/axial reaction, preload, lubrication and assembly procedure |
| Angular-contact arrangement | Combined force and moment support | Contact angle, pair arrangement, preload and moment stiffness |
| Wheel overhang | Known geometry enables reaction calculation | Measure distance from contact/load line to each bearing; check static capacity separately |
4. Validate tread against the actual wheel and floor
Tread evidence: known inputs versus missing ratings
Shore hardness and a polymer brand do not define a universal operating envelope.
| Candidate | Illustrative Cr / μ | Example speed screen | Still unknown |
|---|---|---|---|
| High-resilience PU 95A candidate | 0.012 / 0.68 | 3.5 m/s | Wheel-specific dynamic load, heating and minimum temperature |
| Industrial PU 92A candidate | 0.018 / 0.65 | 2.0 m/s | Formulation, bond system, site friction and duty |
| Soft PU 85A candidate | 0.024 / 0.78 | 1.5 m/s | Deformation, rolling losses, wear and flat spots |
| ESD PU 92A candidate | 0.019 / 0.63 | 2.0 m/s | Electrical resistance of the full wheel-floor path and grip |
The speed screen changes at 75% and 95% of the example reference, or earlier in warm ambient conditions. These are author-selected prompts to seek testing. A low screen value does not establish thermal safety. Below 0°C the tool always requests cold-operation review. Covestro describes Vulkollan® as an NDI-based elastomer with high dynamic capability; its public overview does not support a universal friction coefficient, −40°C approval or speed ceiling. Material source and scope.
5. Keep holding, stopping and scanner configuration distinct
Holding, stopping and sensing need separate evidence
No universal 1.5 multiplier, SIL/PL target, brake delay or optical field size is claimed from public abstracts.
| Function | Tool contribution | Required confirmation |
|---|---|---|
| Static ramp holding | Concept torque at motor shaft | Actual holding torque, drivetrain integrity, wear, tolerance and power-loss tests |
| Dynamic service stopping | Not calculated | Measured worst-case speed/load/grade braking and regeneration limits |
| Emergency stopping | Not calculated | System risk assessment, architecture and validated stop behavior |
| Protective field | Not determined | Installed scanner/control response, stopping distance and scanner-specific supplements |
| Flat-floor parking | Zero ideal grade gives zero modeled grade torque | Slope tolerance, external forces and independent power-off holding strategy |
The holding calculation is 1.5 × m g sin θ × r / (n × ratio). The multiplier is a declared design example; backdrive efficiency, wear and brake tolerances are unresolved. A brake omission on a ramp is flagged, while a nominally level floor still needs a parking strategy. The optical protective field remains undetermined because braking tests, scanner response and installation details are absent. SICK planning instructions; ISO 3691-4 scope.
6. Test heat, lost motion and load transfer
Equal torque sharing gives translational reflected inertia per motor m r² / (n × ratio²); include rotor, wheel and gear inertia separately. Lost motion is an example input by topology, not a supplier promise. The heading illustration assumes two differential wheels, a 0.6 m track and opposing lash take-up; it is not accumulated navigation drift. Real suspension travel requires measured floor profiles, wheelbase, spring rate and load transfer. No generic DIN tolerance establishes a minimum suspension stroke.
Motor, gearbox and inverter losses need a verified heat path and thermal soak test. Constant efficiency cannot establish winding temperature. Bearing electrical erosion, if relevant to the actual drive installation, requires system investigation; no universal switching threshold or bearing replacement interval is asserted here.
7. Compare risks, cost and alternatives
Risk, cost and alternatives
Engineering review agenda. No price, lead-time or durability promise is inferred.
| Risk | Cost / consequence | Mitigation and alternative |
|---|---|---|
| Treating estimates as ratings | Undersized motor, reducer or wheel | Obtain continuous/RMS and peak curves; consider an existing rated module first |
| Traction demand above assumed grip | Slip, failed acceleration or ramp climb | Lengthen acceleration time; verify normal load distribution and measured floor friction |
| Low chassis envelope | CAD rework, tooling and mounting cost | Check wheel diameter and complete CAD; compare motor orientations before custom gearing |
| Cold, wet or rough floor | Seals, tread, corrosion and suspension development | Specify measured conditions and qualification tests; no automatic freezer approval |
| New custom hardware | NRE, tooling, prototypes, spares and validation effort | Request costs and acceptance gates explicitly; standard module may reduce project effort |
| Heat and inverter-related damage | Torque derating, bearing or insulation failure | Thermal soak and system electrical checks; mitigation selected after measurement |
8. Worked concept examples
Hypothetical inputs using the same calculation as the tool. These are not customer installations or test results. Choose the matching preset above to reproduce them.
Under-pallet packaging review
1,200 kg carried payload + 220 kg tare; 1.4 m/s; 2.5% grade; 160 mm wheels. A right-angle candidate changes motor orientation, but does not establish total chassis height.
- Wheel / ratio
- 167.1 rpm / 17.95:1
- Wheel steady / peak
- 20.6 / 109.67 Nm
- DC current, all drives
- 18.7 A
- Assumed grip used
- 39%
Next action: Compare the wheel and motor CAD against the lifting mechanism; obtain actual ratio and continuous torque ratings.
Heavy carried-load AGV
3,500 kg carried payload + 850 kg tare; 1.2 m/s; 5% grade; 250 mm wheels. This is a hypothetical sizing example, not a tested installation.
- Wheel / ratio
- 91.7 rpm / 27.27:1
- Wheel steady / peak
- 181.07 / 424.67 Nm
- DC current, all drives
- 55.9 A
- Assumed grip used
- 35%
Next action: Resolve each bearing reaction and static capacity; measure ramp grip and verify holding under power loss.
High-speed shuttle
350 kg carried payload + 120 kg tare; 2.5 m/s; 0.8 s acceleration; level floor; 160 mm wheels. Speed and load assumptions require wheel and braking validation.
- Wheel / ratio
- 298.4 rpm / 13.4:1
- Wheel steady / peak
- 2.21 / 68.01 Nm
- DC current, all drives
- 3.5 A
- Assumed grip used
- 83%
Next action: Confirm peak motor current and RMS duty, tread thermal behavior, regeneration acceptance and measured worst-case stopping distance.
9. Agree evidence and deliverables before custom manufacture
Agree responsibility for drawings, calculations, fixtures, software interfaces, spare parts and compliance evidence. Development timing, prototype quantity and production lead time depend on the selected design and qualification scope. No response deadline, test-hour guarantee or fixed documentation package is promised.
10. Engineering and procurement FAQ
Sizing and scope
What does custom drivetrain mean on this page?
It means the motor, transmission, hub bearings, holding brake and traction wheel for an AMR or AGV. This page does not size passenger-car powertrains, hydraulic systems or trailer drawbar loads.
How is the reduction ratio calculated?
Wheel rpm equals 60 × speed / (π × wheel diameter in metres). For a geared candidate, target ratio equals entered motor rated rpm / wheel rpm. Select a real available ratio and check the resulting motor speed and torque curve.
Why are wheel and motor torque different?
The reduction increases wheel torque. Estimated motor torque is wheel torque divided by ratio and assumed gear efficiency. Both are per driven unit with equal sharing; real load imbalance needs separate review.
Is steady wheel torque a continuous motor rating?
No. It is the estimated torque for constant-speed travel up the entered grade. Continuous and RMS motor ratings require the complete speed-time-load cycle, thermal environment and supplier curves.
Can this tool size a towing tractor or mecanum vehicle?
The mass model assumes carried payload. Trailer rolling resistance and drawbar transfer require a separate towing model. The tool also omits mecanum roller losses, wheel-force direction and steering scrub; use a topology-specific model for those vehicles.
What happens when traction demand is too high?
The result warns when the peak force exceeds assumed friction times driven normal load. Lengthen the acceleration time, reassess drive-wheel loading or reduce grade/speed. A value below that limit still needs site friction testing.
Components and materials
Does a direct-drive motor run at the entered rated rpm?
Its operating speed equals wheel rpm because the ratio is 1:1. The entered rated rpm is treated as a speed ceiling; the motor must deliver the estimated wheel torque at the actual wheel speed.
Does L10h predict hub service life?
No. The displayed example uses an assumed roller-bearing capacity and simplified radial load. Real bearing reactions, axial loads, preload, contamination and lubrication are unresolved. ISO 281 basic life is statistical fatigue life at 90% reliability, not a warranty.
Is a wheel material approved when the speed screen is normal?
No. The screen compares speed with a declared example reference. It is not a measured temperature or a wheel load/speed rating. Obtain the exact wheel supplier’s continuous and intermittent ratings at your load and ambient temperature.
Can polyurethane hardness establish cold-room suitability?
No. Formulation and the complete wheel construction matter. Confirm low-temperature dynamic performance, bond integrity, seals, lubricant, brake and battery behavior. This page does not assign a universal glass-transition temperature to PU or approve freezer use.
Braking and power
Can the ramp-hold value be used to set a safety laser field?
No. Static holding torque cannot establish dynamic braking. The tool deliberately gives no optical protective field length. Use measured worst-case braking distance, installed scanner/control response times and the scanner manufacturer’s required supplements.
Is the 1.5 holding multiplier an ISO 3691-4 requirement?
It is an author-selected concept multiplier here. The public standard abstract does not establish this value. Confirm applicable complete-vehicle requirements and validate power-off holding with actual components and operating conditions.
Is calculated DC current equal to motor phase current?
No. It is total propulsion DC current estimated from mechanical power and constant efficiencies at speed. Phase current, stall torque, transient peaks, controller limits and auxiliary loads require separate selection.
Does daily Ah estimate determine battery capacity?
No. It estimates only steady propulsion current over the entered operating hours and duty fraction. Add acceleration, auxiliaries, idle, reserve, battery aging and temperature derating, then validate the actual route energy profile.
RFQ and validation
What should be included in an engineering RFQ?
Copy the concept brief and provide the duty cycle, CAD envelope, wheel and floor conditions, braking and parking requirements, target volume and validation criteria. Confirm which drawings, calculations, prototypes and test reports are included in the quotation.
How long does development take and is compliance included?
Timing, NRE, tooling, deliverables and compliance responsibilities need project-specific agreement. The calculator does not commit delivery dates or certify a complete vehicle. Agree acceptance tests before prototype manufacture.
Sources, date and evidence limits
Checked 2026-11-02. Public standard abstracts establish scope; full normative clauses have not been verified here. Engineering equations and author-selected assumptions are distinguished above. Exact product ratings, pricing, lead times and complete-vehicle conformity remain to be confirmed.
- ISO 3691-4:2023 — Driverless industrial trucks and their systems
Public abstract establishes vehicle safety and verification scope. It does not establish the calculator’s coefficients, a universal holding multiplier, scanner response time or a protective field value. Full requirements and applicability need project review.
- ISO 281:2007 — Dynamic load ratings and rating life
Basic bearing rating life uses 90% reliability. The public scope excludes wear, corrosion and electrical erosion. It does not certify a hub arrangement or supply a bearing capacity for this model.
- NSK — Dynamic load ratings and fatigue life
Manufacturer explanation of C/P, the roller-bearing exponent 10/3 and statistical rating life. This supports the illustrative formula, not the assumed C or equivalent load in this calculator.
- ISO 6336-1:2019 — Spur and helical gear load capacity
Reference framework for spur and helical gear load-capacity calculations. It is not a blanket rating method or shock guarantee for every reducer topology.
- Covestro — Vulkollan® polyurethane elastomer
Describes the NDI-based elastomer family and mechanical/dynamic properties qualitatively. This public overview does not establish universal wheel friction, speed, glass-transition or cold-room ratings.
- SICK S300 — Operating instructions, project planning
Scanner-specific instructions (8010948/ZD76, 2025-08-19, section 4.3.7, page 47) account for braking distance, response times and multiple supplements. Use the instructions for the installed scanner; this example manual is not universal.
Bring the estimate to an engineering review
Copy the concept brief, then paste it into the inquiry form with your CAD envelope, route conditions and target volume.