E-scooter thermal-management engineering: IEC 62133-2:2017 § 7.3 thermal abuse, UL 2272:2024 § 21 abnormal-charging + thermal abuse, ISO 12405-4:2018 PEV battery thermal characterization, JEDEC JESD51-1/-2A/-7 R_θJC measurement, IPC-2221A § 6.2 PCB conductor temperature rise, IEC 60068-2-14:2009 thermal cycle Test Na/Nb, IEC 60068-2-30:2005 humidity cyclic Db, ISO 16750-4:2010 thermal/mechanical environmental conditions, MOSFET junction-temperature limit T_J_max 150-175 °C with R_θJC 0.3-2 °C/W (Infineon IPP/IPB series, Onsemi NTMFS, ST STH240N10F7-6), Arrhenius doubling rule (every +10 °C halves component life of NMC/LFP cells), BMS thermal fold-back when T_cell > 45-50 °C (charge cut-off / discharge derate), hub-motor stator copper I²R loss = I² × R_Cu(T) with temperature coefficient α_Cu = 3.93×10⁻³/°C + iron eddy loss P_eddy ∝ B² × f² × t² (Steinmetz), thermal time constant τ_th = R_th × C_th (continuous-vs-peak power derating motor 5-30 s peak / continuous 30-300 s steady-state), TIM (thermal interface materials): Bergquist Gap Pad k=1.5-6 W/(m·K), Arctic MX-6 grease k=8.5 W/(m·K), PCM Honeywell PTM7950 k=8.5 W/(m·K), cooling topologies (natural convection h_nat 5-25 W/(m²·K) / forced air h_forced 25-250 W/(m²·K) / liquid cold-plate h_liquid 500-20 000 W/(m²·K)), thermal-runaway propagation in 18650/21700 cells (T_onset 130-150 °C NMC, 180-200 °C LFP — LFP significantly safer per CPSC + UL data), CPSC recalls (hoverboards 2016 — 501 000 units recalled for thermal runaway, Lime Gen 2 2018 19.2-Wh packs thermal events, Bird Two 2018 charging thermal incidents)
In the guide series we have already covered helmet + protective gear, battery with BMS and thermal-runaway intro, the brake system, motor and controller, suspension, tires, lighting and visibility, frame and fork, display + HMI, SMPS CC/CV charger, connector + wiring harness, IP protection, bearings with ISO 281 L10, stem and folding mechanism, deck, handgrip + lever + throttle, the wheel as an assembly, and bolted-joint engineering as the joining axis. These 18 engineering axes describe individual bricks and the way they are joined — but none of them describes the heat-exchange system that runs through every brick at the same time and requires each component to stay inside its own thermal budget.
An e-scooter is a densely packed thermal system: 600-1500 W of peak power passes through 3-5 energy domains (battery → controller → motor → wheel → road), and each transition dissipates 3-15 % as losses. On a typical 36 V × 15 Ah = 540 Wh pack at 2C discharge (30 A), the pack’s 80 mΩ internal resistance generates 72 W of I²R heat inside the pack itself; a controller with six MOSFETs of R_DS(on) = 5 mΩ contributes another 27 W of switching+conduction loss; a hub-motor with 0.1 Ω phase resistance adds 90 W copper loss + 15-25 W iron loss (Steinmetz). The total ~225 W of thermal power is spread across four locations inside a ~10-15 L volume. Without active or passive thermal management every component’s temperature rises 50-80 °C in 5-15 minutes of continuous full-power riding — and MOSFET T_J_max is 150-175 °C, NMC-cell thermal-runaway onset is 130-150 °C, Class B winding insulation tops out at 130 °C. Those limits are easy to cross in a single full-power hill climb.
This is the nineteenth engineering deep-dive in the guide series — and the second cross-cutting infrastructure axis (parallel to fastener engineering as the joining axis and paired with bearing engineering as the rotation axis + IP engineering as the sealing axis). It describes the way heat is dissipated, which is present in every previous engineering axis: the battery has its own thermal budget; the motor has one; the controller has one; the charger has one. But no component lives in isolation — heat from the motor flows into the frame through the motor mount, heat from the controller flows into the battery via wiring + IP housing, heat from the battery permeates the deck through mounting brackets. All components are coupled through thermal paths — and thermal management consists of ensuring that the sum of heat sources never exceeds the total heat-sink capacity on any temporal horizon (5 seconds for a PWM cycle, 5 minutes for a climb, 1 hour for a journey, 1 year for calendar aging).
CPSC recall history over the last 8 years shows that a substantial share of catastrophic-failure events on e-scooters and adjacent PMD/hoverboards is driven not by mechanical but by thermal mechanisms: Hoverboard recalls 2016 CPSC 16-184 (501 000 units — battery thermal runaway, 99 fires, 18 burn/smoke-inhalation injuries across 24 US states), Lime Gen 2 2018 (battery packs with possible thermal-event scenarios that forced Lime to recall the entire Gen 2 partition of the Bird/Lime fleet), Bird Two 2018 (battery-charging thermal incidents). These are not marginal cases — they are a systematic reminder that thermal management is not optional craft, but a governing-standards discipline (IEC 62133-2:2017, UL 2272:2024, ISO 12405-4:2018, JEDEC JESD51) with quantified requirements.
The scooter owner cannot design the thermal-management subsystem from scratch — but can run an 8-step thermal check and detect 75-85 % of future thermal-event predictors in 90-120 seconds after a ride. That makes thermal engineering the sixth most DIY-accessible engineering axis after bearings, stem, deck/footboard, handgrip/lever/throttle, wheel, and fastener engineering.
Prerequisite: understanding battery engineering (especially the thermal-runaway + BMS sections), motor and controller, SMPS charger, and descending hills + brake thermal management, which treats brake-disc/pad thermal cycles as a separate heat flow.
1. Why thermal management is its own cross-cutting axis
The thermal system is not “it will just cool passively” — it is a system in which every element has quantified engineering specifications:
| Thermal-system element | What it describes | Governing standard |
|---|---|---|
| Heat source | Power dissipation in W, location, temporal profile (PWM / pulse / continuous) | IEC 62133-2:2017 § 7.3 (battery), JEDEC JESD51-1:2012 (semiconductor) |
| Heat path | Material conductivity k [W/(m·K)], cross-section, length, thermal-interface resistance | Fourier’s law Q = k × A × ΔT / L |
| Heat sink | Surface area, fin geometry, convection coefficient h [W/(m²·K)], orientation | Newton’s law of cooling Q = h × A × ΔT, IEC 60068-2-2 |
| Thermal interface material (TIM) | k_TIM, thickness, compression, pump-out resistance | IEC 60068-2-14:2009 thermal cycle, vendor TDS |
| Thermal sensor | Resistance / voltage vs T curve, Beta value, accuracy, response time τ | IEC 60751:2008 (Pt100), JEDEC J-STD-002 |
| Thermal protection | Cut-off / fold-back set-point, hysteresis, response time | UL 2272:2024 § 21.3, IEC 62133-2:2017 |
No element is “standard by default.” A MOSFET in a TO-220 package may have R_θJC anywhere from 0.3 to 2.5 °C/W depending on die size and die-attach quality — at 25 W dissipation, one variant gives Tj = Tcase + 7.5 °C, another gives Tcase + 62.5 °C. The same current through a 4S10P Samsung INR21700-50E pack (50 mΩ × 10 parallel = 5 mΩ × 4S = 20 mΩ pack DCIR) gives 18 W at 30 A; the same pack made from deteriorated Samsung INR18650-29E (100 mΩ × 10 = 10 mΩ × 4S = 40 mΩ pack DCIR) gives 36 W — a 2× difference in thermal power at the same amperage. That makes thermal management its own engineering discipline.
If a MOSFET with R_θJC = 2.5 °C/W is chosen for a spot that expects 30 W continuous dissipation with a sink at 80 °C ambient — Tj = 80 + 30 × 2.5 = 155 °C, which exceeds T_J_max 150 °C for most Si-MOSFETs → solder reflow or die crack in 10-50 hours. This is the analogue of bolt mismatch in fastener engineering (choosing class 4.6 for a motor mount that expects 8.8): geometrically it fits, mechanically it does not; in thermal engineering it fits electrically, but not thermally.
2. Overview of the 8-row standards matrix
E-scooter thermal management is regulated by eight primary standards. Some are product-level safety (UL 2272, IEC 62133-2), others component-level measurement (JEDEC JESD51), still others environmental qualification (IEC 60068-2):
| # | Standard | Edition | Scope | Coverage |
|---|---|---|---|---|
| 1 | IEC 62133-2 | 2017 (+ Amd 1:2021) | Battery cells & packs | § 7.3 thermal abuse: cell heated 5 °C/min to T_max — no fire / explosion; § 7.2.1 short-circuit at high temperature |
| 2 | UL 2272 | 2024 (3rd edition) | Personal e-mobility devices (PMD) | § 21.3 thermal abuse: device-level operation 70 °C ambient × 7 hours; § 21 abnormal charging |
| 3 | ISO 12405-4 | 2018 | Pluggable EV battery packs | § 7.1.6 thermal performance: charge/discharge at -20 to +60 °C; § 7.4 thermal shock |
| 4 | JEDEC JESD51-1 + JESD51-2A + JESD51-7 | 1995 / 2008 / 1999 | Semiconductor thermal measurement | Definition of R_θJC / R_θJA; methodology for still-air natural-convection chamber; test-board geometry |
| 5 | IPC-2221A | 2003 (+ Amd 1:2009) | PCB design | § 6.2 conductor temperature rise: trace-width vs current → 10/20/30/45 °C rise tables |
| 6 | IEC 60068-2-14 | 2009 | Environmental — temperature change | Test Na (rapid change, 2 chambers) + Test Nb (specified rate, single chamber); -55 to +125 °C |
| 7 | IEC 60068-2-30 | 2005 | Environmental — humidity cyclic | Db cyclic test: 25 → 55 °C with RH 95 % cycles over 24 h; condensation on cooled surfaces |
| 8 | ISO 16750-4 | 2010 | Road-vehicle electrical & electronic equipment | § 5.1 thermal storage / cycle; § 5.2 power cycling; § 5.3 thermal shock — applied to e-bike / PMD electronics |
Second-tier standards that support the primary ones: IEC 60751:2008 (Pt100 RTDs), JEDEC J-STD-020E (semiconductor moisture classification), IEC 61010-1:2010 (general electrical equipment safety), MIL-STD-810H Method 501.7 (high temperature) and Method 502.7 (low temperature) — more severe than IEC 60068, used in aviation/military PMD applications.
3. Heat sources on the e-scooter
An e-scooter under continuous full-power operation (for example a 1000-W motor at 25 km/h climbing an 8 % grade) dissipates heat through five localized sources:
| # | Heat source | Continuous power range | Peak power | Mechanism | Location |
|---|---|---|---|---|---|
| 1 | Battery pack I²R + polarization | 15-60 W | 80-200 W (10-s burst at 5C) | DCIR × I² + activation polarization + concentration polarization (Bernardi eq.) | Inside the pack volume; cell-level hot spot in centre cells of the arrangement |
| 2 | Motor controller (MOSFET) switching + conduction | 15-50 W | 60-120 W (during acceleration) | E_sw × f_sw + I² × R_DS(on) per MOSFET × 6 transistors | TO-220 / D²PAK MOSFET family; PCB heatsink area |
| 3 | Hub-motor stator copper (Joule) | 40-150 W | 200-400 W (max-grade climb) | I_phase² × R_phase × [1 + α_Cu(T-25)] × 3 phases | Stator winding inside the rim; thermal hot spot at the slot |
| 4 | Hub-motor iron loss (eddy + hysteresis) | 8-30 W | 15-60 W | k × B^β × f^α × t_lam² (Steinmetz) | Stator-iron lamination |
| 5 | Charger SMPS (only while charging) | 5-30 W | 40-60 W | Switching + transformer winding + diode forward drop | Charger enclosure; transformer core |
Do not confuse this with brake-disc thermal load: brake heat is a separate thermal axis, covered in descending hills + brake thermal management. Brake-disc kinetic-to-thermal conversion (~m × g × h per descent) is a single-event peak (5-30 kJ over 1-2 s), not a steady-state heat source.
Total heat-source power under continuous full-power load: 100-250 W; peak — 300-700 W for 5-10 s. In a pre-warmed device (T_ambient 35 °C, internal temperature 60 °C) critical temperatures of 100-130 °C can be reached in 5-15 minutes of continuous full-power riding — that is the limit that shows up in real-world hill-climbing scenarios and the reason for the derating curves in § 11.
4. Component temperature-limit matrix
Each component category has a quantified maximum; crossing it breaks function or integrity:
| Component | T_max | Mechanism of failure | Reference |
|---|---|---|---|
| NMC 18650/21700 cell | T_onset 130-150 °C (cathode-electrolyte exothermic) | SEI decomposition 60-90 °C → cathode-electrolyte 130-150 °C → thermal-runaway propagation | Tesla/Bosch NMC research + IEC 62133-2:2017 |
| LFP (LiFePO4) cell | T_onset 180-200 °C | Significantly stable cathode (olivine structure); preferred for safety-first applications | UL 2272:2024 + Murata/Sony LFP TDS |
| Si MOSFET (TO-220, D²PAK) | T_J_max 150-175 °C (Si die operating limit) | Die crack, solder reflow, wire-bond lift; AEC-Q101 automotive grade 175 °C | Infineon IPP/IPB, Onsemi NTMFS, ST datasheet |
| NTC thermistor (10K B=3950) | 125-150 °C (operating); 250 °C (storage) | Resistance drift within ±2 % inside the rated range; permanent shift above | Murata NCP15WB / Vishay NTCALUG |
| Electrolytic capacitor (105 °C low-ESR) | 105 °C (rated) → 95 °C (continuous) | Electrolyte vapour pressure → bulge / vent; lifetime doubles per -10 °C (Arrhenius) | Nichicon HW / Rubycon ZL series |
| BLDC stator winding insulation Class B / F / H | 130 / 155 / 180 °C (rated hot-spot) | Varnish breakdown; partial discharge; turn-to-turn short | IEC 60085:2007 thermal classification |
NdFeB rare-earth magnets (rotor in the hub-motor): T_max for N42UH = 180 °C; N48SH = 150 °C; standard N42 = 80 °C — outdoor scooter motors typically use N42SH / N42UH (UH grade specifically for elevated temperature). Curie temperature is ~310 °C, but irreversible demagnetization begins well below the Curie point — typically at 130-160 °C for standard scooter-motor magnets. Exceeding T_max → permanent magnetic-flux loss → motor torque drop with no visible warning to the user (silent failure).
Class B / F / H insulation distinction matters for longevity: operation at rated T_max gives 20 000-hour insulation life (IEEE 1:2000 thermal lifetime); exceeding by 10 °C gives half the lifetime; exceeding by 20 °C gives a quarter. Most scooter motors use Class F (155 °C) as a compromise between cost and durability.
5. Junction temperature and MOSFET R_θJC methodology
Junction temperature T_J is the semiconductor die (silicon crystal) temperature inside the MOSFET package. It is the primary metric for semiconductor reliability — and it cannot be measured directly (the die is packaged). It is calculated through thermal resistance:
T_J = T_C + P_diss × R_θJC (junction relative to case)
T_J = T_A + P_diss × R_θJA (junction relative to ambient — no external heatsink)
Where:
- R_θJC [°C/W] = junction-to-case thermal resistance — measured per JEDEC JESD51-2A (still-air, infinite-heatsink model). Typical TO-220 Si MOSFET: 0.5-2.5 °C/W.
- R_θJA [°C/W] = junction-to-ambient — includes case-to-ambient. Depends on PCB layout, copper-pour area, ambient flow. Typical TO-220 free-air mounted on a 2-oz-Cu PCB: 50-80 °C/W.
Worked example: motor controller with 6× IPB180N04S4-02 (R_θJC = 0.7 °C/W; T_J_max = 175 °C; R_DS(on) = 2 mΩ). Phase current 30 A continuous; PWM 16 kHz at 50 % duty:
P_cond = I² × R_DS(on) × D = 30² × 0.002 × 0.5 = 0.9 W per MOSFET
P_sw ≈ ½ × V_DS × I × (t_r + t_f) × f_sw = 0.5 × 40 × 30 × 50 ns × 16 000 = 0.48 W per MOSFET
P_total = 1.38 W per MOSFET → 8.3 W across all 6 MOSFETs
With Tcase = 70 °C (controller heatsink at mid-load): T_J = 70 + 1.38 × 0.7 = 70.97 °C — comfortable margin. But during a peak acceleration phase of 80 A × 100 ms:
P_cond_peak = 80² × 0.002 × 0.5 = 6.4 W per MOSFET (~5× steady state)
P_sw_peak = 0.5 × 40 × 80 × 50 ns × 16 000 = 1.28 W per MOSFET
P_total_peak = 7.68 W per MOSFET
Transient T_J during a 100 ms burst: T_J(100 ms) = T_C + P × Z_θJ(100 ms) where Z_θJ — transient thermal impedance (typically 0.1-0.3 × R_θJC for a 100 ms pulse) → T_J ≈ 70 + 7.68 × 0.15 = 71.2 °C — still safe. But continuous 80 A → T_J = 70 + 7.68 × 0.7 = 75.4 °C — also safe if the controller heatsink stays at 70 °C in +50 °C ambient. If the heatsink is at 110 °C (degraded TIM, dust-blocked fins): T_J = 110 + 7.68 × 0.7 = 115.4 °C — still under T_J_max 175 °C, but insulation ageing rises exponentially.
Transitive chain: T_J_max → T_case_max (through R_θJC + P) → T_TIM_top_max (through TIM dT) → T_heatsink_max (through TIM-bottom dT) → T_ambient_max. Every link is an R_th element in a Cauer thermal network.
6. Battery thermal management
The lithium-ion battery is the most critical heat source for two reasons: (a) highest energy density (250-300 Wh/kg for NMC) — the largest available thermal energy in case of runaway, and (b) a non-monotonic optimum-temperature window — battery degradation increases both at low temperatures (<10 °C — lithium plating) and high temperatures (>40 °C — SEI + cathode ageing).
Bernardi equation for cell heat generation:
Q_cell = I² × R_internal + I × T × (dV_OC/dT)
└── irreversible Joule ──┘ └── reversible entropy ──┘
The first term is irreversible (always heat); the second is reversible (heat on discharge, cooling on charge for most chemistries; dV_OC/dT ≈ -0.3 mV/K for NMC at SOC 50-80 %).
Arrhenius rate-doubling rule for cell ageing:
k(T) = A × exp(-E_a / (R × T)) (Arrhenius)
For NMC: empirically every +10 °C doubles the calendar-ageing rate (E_a ≈ 30-50 kJ/mol). Translation:
- 25 °C → baseline (1× ageing rate)
- 35 °C → 2× rate (half lifetime)
- 45 °C → 4× rate (quarter lifetime)
- 55 °C → 8× rate
This makes a target operating window of 15-35 °C an absolute imperative for long-life packs. A scooter’s BMS cuts charging at T_cell > 45 °C and derates discharge at > 50 °C — that is thermal fold-back, also covered in battery engineering § BMS.
Thermal-runaway propagation — the catastrophic failure mode where one cell overheats, its heat flows into neighbours, those heat up in turn, and a chain reaction consumes the entire pack:
| Stage | T_cell | Mechanism |
|---|---|---|
| 1. SEI breakdown | 60-90 °C | Solid-electrolyte interphase decomposes, exposing the anode |
| 2. Electrolyte vaporization | 90-120 °C | LiPF6/EC/DMC vapour pressure → swelling |
| 3. Anode-electrolyte reaction | 120-130 °C | Exothermic; CID activates; venting |
| 4. Separator melt | 130-150 °C (PE) / 165 °C (PP/PE/PP trilayer) | Internal short |
| 5. Thermal-runaway onset | 130-150 °C NMC / 180-200 °C LFP | Cathode releases O₂ + heat (>500 °C peak) |
| 6. Propagation to adjacent cells | 200-400 °C | Heat conducts via busbar / case to a neighbour cell at T_onset |
Mitigation: ceramic-coated separators (Al₂O₃) push T_onset 20-50 °C higher; cell-to-cell thermal barriers (aerogel / Pyrogel / mica) slow propagation; cell-holder geometry with air gaps between cells allows venting without heat transfer. LFP chemistry is the best safety-first option (T_onset 50 °C above NMC), but at a density penalty of 30-40 %.
7. Hub-motor: stator copper loss + iron loss + thermal time constant
The hub-motor — a BLDC inside the rim — generates heat through two main mechanisms: copper loss (winding Joule) and iron loss (eddy + hysteresis in the lamination):
Copper loss (temperature-dependent — critical):
P_Cu(T) = I_RMS² × R_phase × [1 + α_Cu × (T - 25 °C)]
Where α_Cu = 3.93 × 10⁻³ /°C is the temperature coefficient of resistance of pure copper. So:
- Phase resistance R_phase = 0.1 Ω at 25 °C
- At 100 °C — R_phase = 0.1 × (1 + 0.00393 × 75) = 0.129 Ω (+29 %)
- At 150 °C — R_phase = 0.1 × (1 + 0.00393 × 125) = 0.149 Ω (+49 %)
This is a positive-feedback loop: hotter winding → higher R → more Joule heat → even hotter. Without active control → runaway in 1-3 minutes of continuous overload.
Iron loss (Steinmetz equation):
P_iron = k × B^β × f^α × t_lam²
Where B is peak flux density (typically 1.0-1.5 T in a scooter motor); f is the electrical frequency (for an 8-pole motor at 1000 RPM = 67 Hz); t_lam is lamination thickness (0.2-0.5 mm for silicon-steel M270); α ≈ 1.5; β ≈ 2.
Iron loss is fixed for a given speed (it does not depend on current/torque) — meaning that at idle or no-load coasting the motor still generates 5-15 W of iron loss (heat without kinetic-energy output). That is why motor warm-up is not just a climbing phenomenon — even steady cruise produces 30-60 W of iron loss.
Thermal time constant τ_th = R_th × C_th:
| Mode | τ_th | Heat budget |
|---|---|---|
| Peak burst (acceleration, 5-30 s) | ~3-10 s (winding only, before heat spreads) | 4-8× rated power tolerable for τ_th × 0.5 |
| Continuous (steady climb, 30-300 s) | ~60-200 s (full motor mass) | Rated power max |
| Steady state (>5 min) | Settled — heat balance reached | Power must be ≤ continuous-rated × derate |
That is why the BMS / controller allows short-duration overcurrent (2-3× current limit for 5-30 s) — it is thermal lag that lets the winding thermal mass act as a buffer before temperature accumulates. Continuous overload is steady-state thermal failure.
8. Charger: thermal fold-back and SMPS efficiency curve
The charger (SMPS — switched-mode power supply) dissipates heat through five sources:
- Bridge-rectifier diode forward drop (4 × 1N5408-class): 4 × 1.2 V × 2 A ≈ 9.6 W at 200 W input
- Switching MOSFET / transistor (D²PAK silicon): conduction + switching loss 5-15 W
- Flyback transformer winding (primary + secondary): copper loss 3-8 W
- Output rectifier diode (Schottky or fast-recovery): 0.5 V × output current ≈ 5-10 W at 5 A
- Output-capacitor ESR ripple: 1-3 W
Total losses 15-50 W at 100-300 W input → η = 80-92 % efficiency typical for a 36 V / 5 A scooter charger.
Thermal fold-back: an internal NTC senses temperature; when T > 60-70 °C the charger reduces output current to keep temperature in check. That is a soft current limit — charge speed drops but the charger does not shut off. If T > 85-90 °C → hard cut-off. Covered in full in SMPS charger engineering § 6.
Constant-current → constant-voltage (CC/CV) thermal characteristic:
- CC phase (0-80 % SOC): full power output → maximum heat
- CV phase (80-100 % SOC): current tapers to ~5 % of rated → losses drop 95 %
So the most thermal stress on a charger is during the CC phase (the first 1-2 hours of a full charge). A charger placed in airflow or on a heat-spreading surface → faster, safer CC phase.
9. Cooling topologies: natural convection vs forced air vs liquid
Three primary modes of heat transfer from source to ambient:
| Mode | h coefficient [W/(m²·K)] | Cost / complexity | Use on a scooter |
|---|---|---|---|
| Natural convection (passive) | 5-25 | Minimum — fin geometry alone | Most commodity scooters; battery pack; controller heatsink |
| Forced air (fan) | 25-250 | Fan + duct + ~2-5 W power | Performance scooters; high-power chargers; some BMSes |
| Liquid cold-plate | 500-20 000 | Pump + coolant + plumbing | Very rare on scooters; common on eMotorcycles / EVs |
| Phase-change cooling (PCM) | Effective ~50-200 (latent-absorption peak) | Material cost only; no moving parts | Some premium battery packs; flagship hub-motors |
Newton’s law of cooling:
Q = h × A × ΔT
Worked example: a heatsink fin area of 0.02 m² (typical TO-247 heatsink), ΔT = 50 °C (sink at 75 °C, ambient 25 °C), natural convection h = 10 W/(m²·K):
Q_max = 10 × 0.02 × 50 = 10 W
So a passively cooled heatsink handles ~10 W continuous at typical scooter ambient. Forced air at h = 100 → 100 W continuous on the same heatsink. That is why performance scooters with 1500+ W controllers almost always include a fan — passive convection is insufficient.
Heat pipes and vapour chambers — passive two-phase devices with effective k of 5000-50 000 W/(m·K) (vs copper’s 401) — spread heat well from source to a larger heatsink area, but they are premium parts rarely seen in scooters under $2000.
10. Thermal interface materials and grease/pad selection
Between a MOSFET / chip and a heatsink there is no ideal contact — surface roughness creates an air gap with k_air = 0.026 W/(m·K) — a terrible insulator. TIM (thermal interface material) fills the gap with k_TIM = 1-15 W/(m·K) — two to three orders of magnitude better:
| TIM type | k [W/(m·K)] | Cure / setup | Pump-out resistance | Typical scooter use |
|---|---|---|---|---|
| Silicone grease (Halnziye HY-883, GD900) | 4-6 | None (paste) | Low (1-3 yr) | Repair / DIY; budget controllers |
| Premium grease (Arctic MX-6, Noctua NT-H2) | 8-9 | None | Medium (3-5 yr) | Enthusiast rebuilds |
| Phase-change material (PCM) (Honeywell PTM7950, Bergquist Hi-Flow) | 5-9 | First heat cycle “wets” the surface | High (5-10 yr) | Premium OEM (Tesla, Bosch) |
| Thermal pad (silicone gap-filler) (Bergquist Gap Pad TGP series) | 1.5-6 | None (compressible) | Excellent (10+ yr) | Battery cell-to-housing; BMS-PCB to enclosure |
| Thermal pad (graphite / PGS) (Panasonic Pyrolytic Graphite Sheet) | 700 (in-plane) / 20 (cross-plane) | None | Excellent (15+ yr) | Heat spreader in tight spaces |
| Thermally conductive epoxy (Henkel Stycast 2850FT, EPO-TEK H20E) | 1-2 (filled) / 30 (silver-filled) | Permanent (hours-days cure) | Permanent | LED-PCB attachment; potted electronics |
Common TIM failure modes:
- Pump-out — repeated thermal cycling makes grease bleed from the central hot zone to the edges → dry spot at the hot zone → spike in R_θCS → MOSFET overheat. Affects cheap silicone-oil pastes the most.
- Dry-out — volatile carrier evaporates above 100 °C ambient → solid powder residue with high R_th.
- Delamination — a silicone pad loses adhesion to the PCB pad after thermal cycling or mechanical vibration.
DIY rule: replace TIM every 3-5 years on a performance scooter; on a budget scooter — after 5-7 years or when performance degrades. Always clean both surfaces with 99 % isopropyl before applying new TIM. Application thickness — for grease 0.05-0.1 mm (just enough to fill); excess raises R_th (TIM has worse k than aluminium / copper itself).
11. Thermal time constants and derating curves
The motor / controller / battery all have non-linear power tolerance as a function of duration and ambient:
6-row derating-curve matrix for motor / controller (typical 1000 W scooter):
| Duration | Ambient 25 °C | Ambient 35 °C | Ambient 45 °C |
|---|---|---|---|
| 5 s peak | 4× rated (4000 W) | 3.5× rated | 2.5× rated |
| 15 s burst | 2.5× rated | 2.2× rated | 1.8× rated |
| 30 s burst | 2× rated | 1.7× rated | 1.4× rated |
| 1 min sustained | 1.5× rated | 1.3× rated | 1.1× rated |
| 5 min sustained | 1.2× rated | 1.0× rated | 0.8× rated |
| Continuous (>15 min) | 1.0× rated | 0.85× rated | 0.7× rated |
That is the practical reason a “1000 W motor” scooter actually has 600-700 W continuous capability at 35 °C ambient — while 1500+ W peak lasts only 5-15 seconds. Marketing-rated power is peak unless explicitly stated; engineering-rated power is continuous at rated ambient.
Battery derating (similar pattern):
- Charge derate at T_cell > 45 °C — current cut to 50 % at 50 °C, full cut at 55 °C
- Discharge derate at T_cell > 50 °C — current cut to 75 % at 55 °C, full cut at 60 °C
- Cold charge cut-off at T_cell < 0 °C — lithium-plating risk
12. Arrhenius rate and component degradation
The Arrhenius equation describes the temperature-dependent rate of any chemically driven degradation:
k(T) = A × exp(-E_a / (R × T))
Where E_a is activation energy [kJ/mol]; R is the gas constant 8.314 J/(mol·K); T is absolute temperature [K]; A is the pre-exponential factor.
+10 °C rule of thumb: for most electronic components and battery chemistries with E_a ~30-60 kJ/mol — rate doubles per +10 °C. Translated to lifetime:
| Component | Rated T | Lifetime at rated T | Lifetime at +10 °C | Lifetime at +20 °C |
|---|---|---|---|---|
| NMC cell calendar ageing | 25 °C | 10 yr (80 % SOH) | 5 yr | 2.5 yr |
| Electrolytic cap (105 °C low-ESR) | 105 °C | 2 000 hours | 1 000 hours | 500 hours |
| Class F motor winding insulation | 155 °C | 20 000 hours | 10 000 hours | 5 000 hours |
| Silicone TIM pump-out | 100 °C | 5 yr | 2.5 yr | 1.25 yr |
Practical implication: keep components 10 °C below rated → 2× lifetime. That is why serious scooter builders oversize heatsinks and use forced air even when passive convection is theoretically sufficient — it is insurance against Arrhenius.
13. 6-row failure-diagnostic matrix
| # | Symptom | Mechanism | What happened | Severity |
|---|---|---|---|---|
| 1 | Cell venting / smoke from the battery enclosure | Thermal-runaway initiation | SEI breakdown → cathode-electrolyte exothermic | Critical — evacuate immediately; class-D fire risk |
| 2 | MOSFET solder reflow / package darkening | T_J > 200 °C transient | Die-attach delamination or solder-pad detachment | High — controller replacement |
| 3 | NTC thermistor drift > ±5 °C | Repeated T_max excursion | Manganese-ion migration; permanent resistance shift | Medium — BMS mis-reading; recalibrate / replace |
| 4 | Electrolytic cap bulge / vent | T > rated 105 °C × extended | Electrolyte vapour pressure → top-vent rupture | High — replace PSU or controller |
| 5 | Hall-sensor drift / phantom signal | T > 125 °C operating | Latch-up or digital trigger error | Medium — motor stalls / cogs; sensor replacement |
| 6 | Stator winding insulation breakdown (smoke / short) | T > Class B/F/H rated | Varnish carbonisation; turn-to-turn short | Critical — motor replacement; potential battery short |
Diagnostic tools:
- K-type thermocouple probe ($10-30) — taped to a MOSFET case / battery exterior; read with a cheap multimeter that has a TC input
- IR thermometer ($20-60) — non-contact spot reading; the emissivity setting matters (default 0.95 for non-shiny surfaces)
- Thermal-imaging camera ($200-1500 entry-level — FLIR C5, Seek Thermal Compact, Fluke TiS20) — the best ROI for serious diagnostics; reveals hot spots invisible to a spot probe
14. 8-step DIY thermal check
| # | Step | What to look for | Tool |
|---|---|---|---|
| 1 | After a 5 min moderate-load ride, park the scooter and briefly touch the top of the battery enclosure | < 40 °C = comfortably warm; 40-50 °C = warm-hot; > 50 °C = check BMS | Finger / IR thermometer |
| 2 | Touch the controller housing | < 50 °C OK; 50-70 °C marginal; > 70 °C = thermal-management issue | Finger / IR thermometer |
| 3 | Touch the hub-motor stator (through the rim) | < 60 °C OK; 60-90 °C high-load expected; > 90 °C = overload | IR thermometer (rim emissivity ~0.3 → adjust!) |
| 4 | Charger surface after 30 min CC charging | < 50 °C OK; 50-65 °C normal; > 65 °C = ventilation issue | IR thermometer |
| 5 | Battery cell-temp readout in the BMS app (if available) | All cells within ±3 °C of each other; max < 45 °C during charge | App / Bluetooth interface |
| 6 | Visual: battery enclosure swelling, melted plastic, discoloration | None | Eyes |
| 7 | Smell: chemical / electrolyte / burning insulation | None — any smell = stop riding | Nose |
| 8 | Thermal-imaging scan (if camera available): controller / battery / motor | Hot spots inside expected zones; no outliers > 20 °C above neighbours | FLIR / Seek / Fluke |
Run this check after every ride longer than 5 km on performance scooters; after rides > 15 km or > 30 °C ambient on commuter scooters. Halt and investigate at any sign of distress.
15. 6-step DIY remediation
| # | Issue found | DIY-doable | Action |
|---|---|---|---|
| 1 | Battery > 50 °C after a moderate ride | Yes | Park in shade; let it cool; check BMS app for cell imbalance; reduce load |
| 2 | Controller > 70 °C | Yes (if accessible) | Open enclosure; clean dust from heatsink fins; replace TIM if dry or cracked |
| 3 | Hub-motor > 100 °C | Partially | Reduce continuous load; check wheel drag (bearings, tire pressure, alignment); avoid sustained climbs |
| 4 | Charger > 65 °C | Yes | Move to a well-ventilated location; do not charge on a carpet / blanket / bedside; check that vents are clear |
| 5 | Cell imbalance (>50 mV between cells at rest, >100 mV under load) | No (DIY rebalance is risky) | Take to a qualified e-scooter shop; balanced charge with lab equipment |
| 6 | Stator winding smell / smoke | No | End of life — replace motor; possible battery damage; STOP USING |
16. Case studies — CPSC and industry incidents
Case 1: Hoverboard recalls 2016 (CPSC 16-184) — 501 000 units, 8 distinct importers (Swagway, Razor Hovertrax, Hoverboard LLC, Powerboard, etc.). Mechanism: low-quality 18650 cells without UL 2272 certification (which wasn’t yet mandatory) in packs with inadequate thermal management — cells went into thermal runaway during/after charging; fires reported in 24 US states; 99 fire incidents, 18 burn injuries, $2.5 M property damage. Root cause: counterfeit / mislabelled NMC cells with internal defects; pack designs without thermal barriers between cells; chargers without proper end-of-charge thermal monitoring. Outcome: this catalysed the creation of UL 2272 (2016 first edition; current 3rd edition 2024) — now mandatory for PMD in US / CA / UK / AU.
Case 2: Lime Gen 2 thermal events 2018-2019 — Lime voluntarily recalled the Gen 2 fleet after battery thermal events in multiple US cities. Mechanism: the battery enclosure did not dissipate heat fast enough under intense summer use (Phoenix / Austin / Dallas — 40+ °C ambient); BMS thermal-fold-back set points did not account for cumulative thermal stress after sustained 12+ hour fleet operation. Outcome: Gen 3 and later Lime/Bird models use automotive-grade battery management with cell-level thermal sensing and active cooling in some regional fleets.
Case 3: Bird Two charging thermal incidents 2018-2019 — Bird voluntarily replaced multiple Bird Two units after reports of battery-pack thermal events during charging at warehouses. Mechanism: chargers operating at high ambient (warehouses without HVAC in Texas) with many chargers packed close together → cumulative heat load on the shared environment → individual chargers operating at their upper thermal limit → occasional thermal cut-off failures. Outcome: Bird (and the industry as a whole) introduced ventilation standards for charging facilities, charger duty-cycle limits, and visual / smoke detectors in all charging warehouses.
Industry-response trend: after 2020 the PMD industry shifted increasingly toward LFP chemistry (vs NMC) for shared-fleet applications. LFP has lower energy density (slightly more mass + volume per kWh) but thermal-runaway onset of 180-200 °C vs NMC’s 130-150 °C — significantly safer for charging facilities and high-temperature environments. Premium personal scooters still lean toward NMC for range / weight, but flagship models increasingly include cell-level temperature sensors and ceramic separator coatings.
17. Recap — 10 key takeaways
-
Thermal management is a cross-cutting infrastructure axis parallel to fastener (joining) / bearing (rotation) / IP (sealing) → thermal = heat-dissipation axis. It does not describe a specific component; it describes the way every previous component receives and rejects heat.
-
Heat sources on a scooter — 5 sources: (1) battery I²R + polarization (15-60 W); (2) controller MOSFET switching + conduction (15-50 W); (3) hub-motor copper I²R (40-150 W); (4) hub-motor iron loss (8-30 W); (5) charger SMPS (5-30 W while charging). Total 100-250 W continuous, 300-700 W peak.
-
Component limits: NMC T_onset 130-150 °C; LFP T_onset 180-200 °C; Si MOSFET T_J_max 150-175 °C; Class F winding 155 °C; electrolytic cap 105 °C; NdFeB N42UH magnet 180 °C (irreversible demag at 130-160 °C).
-
Junction temperature T_J via R_θJC methodology (JEDEC JESD51-2A): T_J = T_C + P × R_θJC. Transitive chain T_J → T_TIM → T_heatsink → T_ambient through a Cauer thermal network of R_th elements.
-
Battery thermal management — Bernardi equation (Joule + entropy); Arrhenius +10 °C rule (ageing rate doubles per 10 °C); BMS thermal fold-back at T_cell > 45 °C; thermal-runaway propagation through 6 stages (SEI → electrolyte vap → anode-electrolyte reaction → separator melt → runaway onset → propagation). Ceramic separators and cell-to-cell aerogel barriers — mitigations.
-
Hub-motor thermal: copper loss with positive feedback (α_Cu = 3.93 × 10⁻³ /°C); Steinmetz iron loss P = k × B^β × f^α × t_lam² (fixed for given speed, independent of current); thermal time constant τ_th 60-200 s (continuous-rated power) vs 3-10 s (winding-only peak burst — allows 2-4× rated power for 5-30 s).
-
Charger thermal: SMPS efficiency 80-92 %; 5 heat sources (rectifier diodes, switching MOSFET, transformer winding, output diode, output-cap ESR); CC phase = max heat (first 1-2 hours); thermal fold-back at > 60-70 °C reduces output current.
-
Cooling topologies: natural convection h 5-25 W/(m²·K) — most commodity scooters; forced air h 25-250 — performance and premium; liquid cold-plate h 500-20 000 — eMotorcycles / EVs (rare on scooters). TIM selection — silicone grease (4-9 W/(m·K), pump-out 1-3 yr); PCM (5-9, 5-10 yr); thermal pad (1.5-6, 10+ yr); thermally conductive epoxy (1-30 silver-filled, permanent).
-
Derating curves — power tolerance is non-linear in duration and ambient: 4× rated for a 5 s peak at 25 °C, 1× continuous at 25 °C, 0.7× continuous at 45 °C ambient. Arrhenius rule — operating 10 °C below rated → 2× lifetime; that is why serious builders oversize heatsinks.
-
DIY check — 8 steps after every 5+ km ride: battery, controller, motor, charger temperatures; visual signs (swelling, discolouration); smell (electrolyte, insulation burn); BMS app cell readings; thermal-imaging scan if a camera is available. CPSC case studies: hoverboards 2016 (501 000 units, catalysed UL 2272); Lime Gen 2 2018-2019 thermal events; Bird Two charging-facility incidents. Industry-wide shift toward LFP chemistry for shared fleets — a safer thermal profile at the cost of slightly higher mass / volume.