Robot Colaborativ (Cobot) — OEM China

ISO/TS 15066 și Directiva CE pentru Mașini: Ce Acoperă de Fapt Certificarea

ISO/TS 15066 defines the power-and-force limits for collaborative robot operation — the technical standard behind every “safe for human-robot collaboration” claim. Understanding what it covers (and what it does not) is essential before committing to a supplier.

Body region contact force limits (ISO/TS 15066, Table A.2): The standard specifies transient contact force limits by body region. Relevant thresholds for typical applications: head and neck (130N), sternum (140N), finger (140N), palm (180N), forearm (180N). These are not robot design specs — they are the limits the complete robot cell must not exceed during an unintended contact event. The cobot manufacturer’s job is to provide a robot capable of being configured to stay within these limits. The integrator’s job is to verify the complete cell meets them.

Speed-separation monitoring (SSM) is not built into the robot. SSM — where the robot reduces speed as a human enters a monitored zone — requires a validated external safety sensor: laser scanner (SICK S3000, Pilz PSENscan), camera-based system (Pilz PITreader), or lidar. The cobot itself does not provide SSM. If a Chinese supplier quotes “SSM capable” as a robot feature, ask specifically whether a safety-rated sensor is included and which EN/IEC standard it is certified to. Usually it is not included.

CE marking under EU Machinery Directive 2006/42/EC: A CE-marked Chinese cobot means the manufacturer has issued a Declaration of Conformity and compiled a Technical File for the robot as an isolated machine. This satisfies the robot’s own CE obligations. It does not certify the complete robot cell — the end-user assembly of cobot + end-effector + workpiece fixture + safety zone — as safe. Under the Machinery Directive, the complete robot cell is itself a “machine” and requires its own risk assessment and CE mark per EN ISO 10218-2. A common mistake from Western buyers: assuming the cobot CE mark covers the installation. It does not. Budget $5,000–20,000 for a qualified integrator to complete the risk assessment and cell-level CE documentation.

Practical implication: When evaluating Chinese cobot suppliers, request the Declaration of Conformity and the test reports behind it — specifically which notified body (TÜV SÜD, TÜV Rheinland, Dekra, Bureau Veritas) issued the ISO 10218-1 type examination certificate. Factory self-declarations without a notified body certificate are insufficient for EU market deployment. Our sourcing service pre-screens suppliers for third-party certificate validity as a baseline qualification step.

Detectarea Cuplului Articulației vs Monitorizarea Viteză-și-Separare: Compromisul Cost-Siguranță

The two primary methods for achieving collaborative operation have meaningfully different cost and integration implications. Chinese cobots below approximately $15,000 typically omit joint torque sensors — understanding the implications prevents costly surprises post-purchase.

Joint torque sensing (JTS) — direct collision detection. Each joint contains a torque sensor. When the robot detects an unexpected torque spike (contact with a person or object), it halts within <10ms. This enables Power and Force Limiting (PFL) operation without any external sensor — the robot itself detects and responds to contact. JTS is required for true PFL collaborative operation per ISO/TS 15066. Cobots with JTS: AUBO i-series (external torque sensor option), Han’s Robot Elfin-series, Rokae CR-series. JTS adds $3,000–8,000 to unit cost depending on payload class.

Speed-and-separation monitoring (SSM) — pre-collision zone detection. The robot uses an external validated safety sensor to detect human presence and reduce speed before contact occurs. This avoids contact entirely (at the cost of a permanently slower cycle time in shared zones) and does not require JTS. SSM is the approach used by most entry-tier Chinese cobots: JAKA Zu series base configuration, Elite Robots EC series, Dobot CR series. The robot does not have independent collision detection capability — it relies entirely on the external safety system.

How to verify which type you are buying: Request the PLd or PLe safety function certificate specifically for the collision detection mechanism. A PLd Cat.3 certificate for “robot cell with external SSM scanner” is a different product than a PLd Cat.3 certificate for “robot with integral joint torque sensing.” Both can be legitimate, but they have different integration requirements and different residual risks. Do not accept a single CE mark as proof — get the safety function certificate from the notified body that shows which collision detection method was evaluated.

Selection guidance: For applications where the robot and operator regularly share workspace simultaneously (assembly assistance, machine tending with hand-offs), JTS is the correct choice. For applications where humans and the robot occupy the zone at different times (operator loads part → steps back → robot cycles), SSM with a validated safety scanner is sufficient and significantly cheaper. Our inspection service includes verification of the safety certificate chain before shipment.

Repetabilitatea, Precizia Traiectoriei și Costul Total de Deținere

Chinese cobot datasheets reliably quote repeatability. They rarely quote path accuracy, thermal drift, or accuracy at payload extension — three numbers that matter more in production than the headline repeatability figure.

Repeatability vs path accuracy. ISO 9283 repeatability: ±0.03mm means the robot returns to the same point within ±0.03mm across 30 repeated cycles, measured at workspace center, at rated payload, at operating temperature. This is the number on every Chinese cobot datasheet. Path accuracy (also defined in ISO 9283) measures how closely the robot follows a commanded Cartesian path between two points — relevant for welding seam quality, glue bead consistency, and deburring. Chinese suppliers frequently omit path accuracy data. If your application involves continuous path motion (welding, dispensing, laser cutting), request the ISO 9283 path accuracy test report specifically. Expect path accuracy of 0.15–0.5mm for Chinese cobots in the $10,000–25,000 range — adequate for most assembly, inadequate for precision welding without seam tracking.

Accuracy at full extension, full payload. The ±0.03mm spec is measured at workspace center (roughly 50–60% of maximum reach) with rated payload. At full extension with full payload, joint flex and link compliance reduce effective repeatability to ±0.08–0.15mm in most Chinese cobots. For practical applications at 60% of maximum reach with 50% of rated payload, expect to see ±0.03–0.05mm — close to the spec. Size your cobot for the actual working envelope, not the maximum reach spec.

Total cost of ownership (TCO) for a single cobot cell:

Component	Typical Range
Cobot unit (10kg payload class)	$12,000–28,000
End-effector (gripper, welding torch, etc.)	$1,000–8,000
Safety assessment and integration	$5,000–20,000
Programming and commissioning (40–120 hours)	$4,000–18,000 (local integrator rate)
Total cell cost	$22,000–74,000

The cobot hardware is 40–60% of total cell cost in most installations. Buyers optimizing only on robot unit price frequently underestimate integration cost. Chinese cobots at $8,000–15,000 can represent genuine value, but require a competent local integrator — factor $150–250/hour for Western Europe or North America for programming and commissioning.

Chinese cobot ROI vs industrial robot break-even. A comparable 6-axis industrial robot (Fanuc M-10, ABB IRB 1200) costs $18,000–35,000 for the arm alone but requires guarding ($8,000–25,000 in steel safety fence, interlocks, and light curtains), adds 3–6 square meters of floor space, and requires specialist programming. For production volumes above 50,000 cycles/year, the industrial robot typically has lower unit-cost depreciation — cycle times are faster and the arm lasts 80,000+ hours vs 30,000–50,000 hours for most Chinese cobots. For applications with frequent changeover (batch sizes below 500 units, 3–10 product variants), the cobot’s lower integration cost and faster reprogramming make it the economically correct choice regardless of unit price.

Peisajul Furnizorilor Chinezi de Roboți Colaborativi

The Chinese cobot market has consolidated around five credible manufacturers for Western OEM buyers, each with a different position on the quality-cost curve.

Established Chinese manufacturers (qualified for Western OEM):

• AUBO Robotics (Beijing) — 3–20kg payload range, ISO 10218-1 certified via TÜV SÜD, ROS2 SDK actively maintained, European distribution through AUBO Europe. Pricing: $14,000–35,000 depending on configuration. The most documentably mature Chinese cobot for Western integration.
• JAKA Robotics (Shanghai) — 3–18kg payload, competitive pricing ($10,000–28,000), growing European footprint. Entry-tier models use SSM rather than JTS — verify configuration carefully. Python SDK and web-based programming interface are well-documented.
• Elite Robots (Changzhou) — EC66 and EC612 popular in light assembly. CE marked, TÜV Rheinland certificate. Pricing: $9,000–22,000. Documentation quality in English is adequate; after-sales support in Europe via Elite’s German subsidiary.
• Han’s Robot (Shenzhen, subsidiary of Han’s Laser) — Elfin series, 3–15kg. Integrated with Han’s Laser manufacturing ecosystem. Torque sensing standard on mid/high-tier. Pricing: $12,000–32,000.
• Rokae (Beijing) — CR-series, 3–18kg. Targets the mid-tier industrial segment, stronger on path accuracy than assembly-only competitors. Less established European support network.

Quality indicators to verify before committing to a supplier:

1. ISO 10218-1 certificate from an accredited CB. The certificate must name a notified body (TÜV, Dekra, Bureau Veritas, SGS) and include the certificate number. Factory self-declarations that cite “ISO 10218-1 compliance” without a certificate number are not equivalent. Our audit service confirms certificate validity directly with the issuing body.

2. ISO 9283 test report at rated payload. Not a datasheet — a lab test report showing repeatability measurements at rated payload, workspace center, operating temperature, with at least 30 cycle measurements per direction. If the supplier cannot produce this report for the specific model, the spec is marketing copy.

3. SDK documentation and GitHub activity. For ROS2 and Python SDK: check the GitHub repository directly (most Chinese cobots have public repos). Commit frequency, issue response time, and English documentation quality are reliable proxies for long-term software support quality. A cobot with an unmaintained SDK becomes difficult to integrate with evolving robot cell software stacks.

4. European after-sales support presence. For Western buyers, warranty and spare parts logistics matter. AUBO Europe (Germany), Elite’s German subsidiary, and Han’s Robot’s European presence are the current benchmarks. JAKA’s European support is improving but remains thinner than AUBO’s. Confirm: lead time for replacement joint motor, availability of local service engineers, warranty response SLA.

Universal Robots (Danish) as benchmark. UR10e (10kg, $40,000–50,000 for arm only) remains the reliability and ecosystem benchmark — larger integrator network, better documented safety cases, longer field deployment history. For buyers entering the Chinese cobot market: the total cell cost gap between a Chinese cobot + local integration and a UR + local integration is often $8,000–18,000. Whether that gap justifies the ecosystem and support differential depends on production volume, reprogramming frequency, and local integrator UR experience. For applications where an experienced UR integrator is available locally, the Chinese cobot cost advantage can be partially offset by longer commissioning time.

For OEM sourcing, supplier selection, pre-shipment inspection, and certificate verification for collaborative robots, contact us via the industrial IoT sourcing page or submit a project brief through our sourcing service.