Motors · Bafang
Bafang · Bafang M-Series (light e-MTB / e-road / e-gravel)

M820

Bafang's lightest mid-drive, the M820 chases the lightweight-eMTB class with a 2.3kg magnesium drive unit and a torque split that independent testers read as roughly 75Nm sustained and a peak near 95Nm, topped by a 'Rocky' boost mode whose real-world value reviewers question.

M820 eMTB motor
The Bafang M820 drive unit — 2.3kg, magnesium-cased, rated 250W. Bafang lists 80Nm 'Max. Possible torque'; independents read it as ~75Nm sustained / ~95Nm peak. (Bafang)
0250500406080100120610 Wcadence (rpm) →power (W)

Representative shape (not a measured dyno trace): builds quickly off the bottom, peaks in the mid-cadence band around 70-90rpm, then tapers smoothly as it spins toward its 120rpm ceiling. Voltriderz puts achievable peak output near 600W, but that figure is tied to the 48V build and updated firmware — the 43V OEM units most riders run sit lower.

The verdict

Bafang M820 is the Chinese giant's bid for the lightweight trail class, and on paper it lands in the gap between the featherweights and the full-power units. At 2.3kg it undercuts a Bosch CX by the best part of a kilo. Bafang's own page lists a single headline figure of 80Nm, which it labels 'Max. Possible torque' rather than a sustained rating; independent write-ups (ebike24, Voltriderz) read that as roughly 75Nm effective with a peak nearer 95Nm. Either way it sits clear of the 60Nm Fazua Ride 60 and Bosch SX it competes with.

The case is magnesium, which Bafang leans on for heat management and vibration damping rather than outright stiffness, and reviewers credit it with holding output on long climbs where cheaper alloy housings begin to throttle. Drive is metered by a torque sensor, so it reads pedal pressure cleanly and doesn't surge; the trade-off is that you have to put real watts in to unlock the full assist. Max cadence stretches to 120rpm, so it spins out later than the slower light motors. The headline trick is the 'Rocky' boost mode, but it's worth tempering expectations: Voltriderz's tester found it 'feels more like a marketing gimmick than a practical feature', with a boost too brief and too fiddly to call up on a real technical climb. Treat the ~95Nm peak as a short, conditional burst, not a mode you ride in.

No independent dyno bench has published a measured power curve for the M820, so the figures here are Bafang's claims plus a representative measured-shape curve, and the sources are the manufacturer page and reviewer reports rather than lab data. Voltriderz reckons around 600W is achievable, but only on the 48V build with the right firmware: the motor ships in 43V (OEM-dominant), 36V and 48V variants, and the 43V units most riders get won't see that number. As a value-tier route into a sub-18kg eMTB it's compelling on price and weight; what it still lacks is the long independent test record that Bosch, Shimano and TQ can point to.

“Roughly 75Nm sustained in a 2.3kg case, peaking near 95Nm — the value route into a sub-18kg eMTB, if you can live without a bench-test record.”

Character

Rider input
Bafang publishes no support-ratio figure for the M820, so there's no headline assist multiplier to quote. Delivery is torque-sensor metered — it mirrors pedal pressure, so you push real watts to call up the full output rather than ghost-pedalling for it.
On the trail
Natural, pressure-following delivery rather than an off-the-line shove; the 'Rocky' mode gives a short extra kick for steep pitches — though testers find it too brief and fiddly to lean on — and the 120rpm ceiling keeps it pulling when lighter motors have spun out.
Noise
Mid-volume whirr under load — quiet for the lightweight class but not silent. Voltriderz also notes a distinctive rattle on descents and coasting that undermines the premium feel. No reviewer has published a measured dBA figure for the M820, so the comparison with the quieter TQ HPR50 is qualitative only.
Efficiency
Low Q-factor and a clean torque-sensor tune help it turn pedal input into drive efficiently, and reviewers rate real-world efficiency as good. No measured efficiency, range or Wh/km figures have been published for the M820, and host-bike range claims (e.g. ~90km on the CEF50's 720Wh pack) are manufacturer estimates rather than tested numbers.

The case for and against

Strengths

  • Only 2.3kg — enables sub-18kg complete builds
  • ~75Nm sustained, ahead of 60Nm light-class rivals, peaking near 95Nm
  • Magnesium case manages heat well on long climbs
  • 120rpm ceiling resists spin-out
  • Cheap route into a light, full-carbon eMTB via open-mould frames

Compromises

  • No independent dyno, thermal or dBA bench data — specs are claims, not measured
  • 'Rocky' boost rated a marketing gimmick by reviewers — too brief and fiddly to use
  • ~600W peak only on the 48V firmware build; 43V OEM units sit lower
  • Distinctive rattle on descents — not as quiet as a TQ HPR50
  • Thin field record on emerging open-mould brands; no major-brand support network

How it stacks up

It sits above the 60Nm Fazua Ride 60 and Bosch SX on torque while staying close on weight, but it doesn't reach the 85Nm-plus, full-power Bosch CX or Shimano EP801 class. Against the TQ HPR50 — its closest named rival in the quiet-and-light bracket — the M820 is heavier and, by ear, busier: testers report a descent rattle the TQ doesn't have, and the HPR50 is the benchmark for a near-silent harmonic-pin drive. Neither has a published dBA figure, so the noise gap is described, not measured. Where the M820 hits back is price and outright torque; the real gap to Bosch, Shimano and TQ is the long independent test record, not the spec sheet.
OWNER INTELLIGENCE
Most owners report no problems.
Back
Top