---
title: "Xiaomi CyberGear specs"
description: "QDD / Proprioceptive actuator from Xiaomi (China). Peak torque 12 Nm, rated 4 Nm, weight 317 g, torque density 37.9 Nm/kg, price $179."
canonical: https://actuatorlist.com/actuators/xiaomi-cybergear
last-updated: 2026-11-01T14:05:43.000Z
---
# CyberGear Micromotor by Xiaomi

> QDD / Proprioceptive actuator from Xiaomi (China). Peak torque 12 Nm, rated 4 Nm, weight 317 g, torque density 37.9 Nm/kg, price $179.

## Specifications

| Spec | Value |
| --- | --- |
| Category | QDD / Proprioceptive |
| Peak torque | 12 Nm |
| Rated torque | 4 Nm |
| Max speed | 296 rpm |
| Gear ratio | 7.75:1 |
| Weight | 317 g |
| Torque density | 37.9 Nm/kg |
| Price | $179 |
| Price per Nm | $14.9 |
| Backdrivable | yes |
| Backlash | n/a |
| Voltage | 24 V |
| Peak current | n/a |
| Communication | CAN |
| Control modes | position, velocity, torque, mit |
| Encoder | 14-bit |
| Integrated driver | yes |
| Availability | eol |
| Data status | Catalog headline specs spot-checked: confirm revision |

## Size a joint

- [Motor torque calculator](/tools/motor-torque-calculator?solve=power&t=4&n=296): **124 W** of shaft power at its rated 4 N·m and 296 rpm.
- [Robot arm torque calculator](/tools/robot-arm-torque-calculator?L=0.4,0.35,0.12&m=1.2,0.8,0.3&a=0.317,0.5,0&p=2): **31.2 N·m** at the shoulder with its 317 g at the elbow of a 2 kg-payload arm.

## Use cases

- robot arm
- quadruped leg
- legged robot

## Notes

Xiaomi's low-cost MIT-Cheetah-style micromotor (from the CyberDog / CyberOne ecosystem), popular for hobby legs and arms. 12 Nm peak / 4 Nm continuous, 7.75:1 reduction, 24V, max 296 rpm, 14-bit encoder, CAN (1 Mbps) plus serial, integrated FOC driver with position/velocity/torque/impedance (MIT) modes. Manufacturer-stated torque density 37.85 Nm/kg. Listed EOL by some distributors. Price $179 is the AiFitLab price for the bare motor (no cables or USB-CAN adapter), checked Oct 2026; launch price was CNY 499.

## Engineering comparison

Read the [AK80-9 V3.0 KV100 versus original CyberGear comparison](/comparisons/cubemars-ak80-9-vs-xiaomi-cybergear) for source-qualified efficiency, thermal duty, precision and modeling boundaries. Hardware/firmware identity and unresolved document conflicts matter; this is not a matched bench test. See the [selection guide](/guides/robot-actuator-selection) for a proposed validation protocol.

## Engineering evidence

**Revision scope:** Original CyberGear, 24 V / 7.75:1 / 317 g; not CyberGear 2.0

- Manufacturer-branded Chinese manual retrieved from an unauthenticated third-party mirror with expired TLS certificate; provenance confidence moderate. Installed hardware/firmware unpinned. Document review, not bench validation.

### Efficiency

**Evidence: unknown.** No complete bus-to-output efficiency map established in the reviewed original manual; its torque–speed curve is not efficiency.

**Scope / unknowns:** Iq/phase current and bus voltage do not substitute for synchronized bus-power and output-torque measurements; regeneration and idle losses remain unknown.

**Source locators:**

- [Chinese manual (mirror)](https://roboticscv.com/wp-content/uploads/pdf/CyberGear%E5%BE%AE%E7%94%B5%E6%9C%BA%E4%BD%BF%E7%94%A8%E8%AF%B4%E6%98%8E%E4%B9%A6.pdf): PDF pp4,6,15–16; parameter/telemetry tables; reviewed pp1–30

### Thermal duty

**Evidence: manufacturer.** Vendor overload table includes 11 N·m/45 s, 10/60, 9/90, 7/320 at 24 rpm, 25 °C ambient and 120 °C winding endpoint.

**Scope / unknowns:** 120 °C is not a protection setting: board ceiling 80 °C, CAN default motor warning/fault 75/80 °C. Initial winding temperature, mounting/airflow, NTC-to-winding offset and repeat-cycle recovery are unknown; no cooldown inferred.

**Source locators:**

- [Chinese manual (mirror)](https://roboticscv.com/wp-content/uploads/pdf/CyberGear%E5%BE%AE%E7%94%B5%E6%9C%BA%E4%BD%BF%E7%94%A8%E8%AF%B4%E6%98%8E%E4%B9%A6.pdf): PDF pp4–5 section 1.3.15 overload table; p6 board limit; p24 section 4.1.10 CAN defaults

### Precision

**Evidence: unknown.** 14-bit single-turn sensing and reported load-side position do not establish loaded accuracy, mechanical play or bidirectional repeatability.

**Scope / unknowns:** Independent external-reference loaded metrology was not established in these reviewed sources; resolution is not accuracy.

**Source locators:**

- [Chinese manual (mirror)](https://roboticscv.com/wp-content/uploads/pdf/CyberGear%E5%BE%AE%E7%94%B5%E6%9C%BA%E4%BD%BF%E7%94%A8%E8%AF%B4%E6%98%8E%E4%B9%A6.pdf): PDF pp6,10,15–16,25; encoder and position feedback tables

### Modeling

**Evidence: research.** Author-reported exploratory MuJoCo identification fits friction/inertia, then manually retunes inertia after a 10 Hz mismatch, worsening step-response agreement.

**Scope / unknowns:** Not peer-reviewed or a portable calibrated model: torque scaling assumed, revision unpinned, raw data and held-out load/temperature errors unavailable. Record AppCodeVersion/AppGitVersion and actual gains; CAN reads 0x7019–0x701D require unspecified newer firmware and are not debugger-only extended telemetry. Type18 writes are power-off volatile.

**Source locators:**

- [hossyan exploratory identification](https://zenn.dev/hossyan/articles/2406675c7d76a3): Sections 2.1–2.3/results/high-frequency validation; manual inertia retuning
- [Chinese manual (mirror)](https://roboticscv.com/wp-content/uploads/pdf/CyberGear%E5%BE%AE%E7%94%B5%E6%9C%BA%E4%BD%BF%E7%94%A8%E8%AF%B4%E6%98%8E%E4%B9%A6.pdf): PDF p13 identity; pp14–16 debugger fields; pp24–25 volatile writes/conditional CAN reads

## Sources

- [https://openelab.io/products/xiaomi-cybergear-micromotor-intelligent-motor](https://openelab.io/products/xiaomi-cybergear-micromotor-intelligent-motor)
- [https://github.com/belovictor/cybergear-docs/blob/main/instructionmanual/instructionmanual.md](https://github.com/belovictor/cybergear-docs/blob/main/instructionmanual/instructionmanual.md)
- [https://hackaday.io/project/202366-the-unnamed-spider-bot/log/239239-steadywin-gim6010-8-vs-xiaomi-cybergear-micromotor](https://hackaday.io/project/202366-the-unnamed-spider-bot/log/239239-steadywin-gim6010-8-vs-xiaomi-cybergear-micromotor)
- [https://aifitlab.com/products/xiaomi-cybergear-micromotor](https://aifitlab.com/products/xiaomi-cybergear-micromotor)

## Similar QDD / Proprioceptive actuators

- [GIM8108-8 (Steadywin)](https://actuatorlist.com/actuators/steadywin-gim8108-8): 58.7 Nm/kg
- [AK10-9 (V3.0 KV60) (CubeMars (T-Motor))](https://actuatorlist.com/actuators/cubemars-ak10-9): 55.2 Nm/kg
- [AK70-10 (KV100) (CubeMars (T-Motor))](https://actuatorlist.com/actuators/cubemars-ak70-10): 47.6 Nm/kg
- [AK80-9 (V3.0 KV100) (CubeMars (T-Motor))](https://actuatorlist.com/actuators/cubemars-ak80-9): 45.4 Nm/kg
- [GO-M8010-6 (Unitree)](https://actuatorlist.com/actuators/unitree-go-m8010-6): 44.7 Nm/kg

## Catalog comparisons

Catalog specifications and source caveats, not compatibility recommendations.

- [AK60-6 (V3.0 KV80) vs CyberGear Micromotor](/comparisons/cubemars-ak60-6-vs-xiaomi-cybergear)
- [qdd100 beta 3 vs CyberGear Micromotor](/comparisons/mjbots-qdd100-vs-xiaomi-cybergear)
- [GIM6010-8 vs CyberGear Micromotor](/comparisons/steadywin-gim6010-8-vs-xiaomi-cybergear)

[Browse actuator comparisons](/compare).

## Machine-readable

- JSON: https://actuatorlist.com/api/v1/actuators/xiaomi-cybergear
- Catalog index for agents: https://actuatorlist.com/llms.txt
