0.2mm of Drift, One Rejected Order: How to Verify a Used Robot's Real Repeatability

Figure 1. Measuring repeatability in the test bay: the number on the report should match the robot on the floor.

A buyer in Southeast Asia bought a 'refurbished' robot from a trader who had cleaned the paint and photographed it well. The machine looked perfect in the photos. On the line, it could not repeat a position within tolerance: joints drifted by more than 0.2 mm on a job that required 0.1 mm, and the first batch of parts was rejected. The trader had never measured precision, because measuring it would have shown what the paint was hiding. That order of rejected parts is what a real inspection prevents.

The one number that matters most

Repeatability is the ability of the robot to return to the same taught point over and over. The Chinese standard GB/T 12642-2013, which mirrors international practice, puts typical six-axis repeatability at plus or minus 0.08 mm or better, with collaborative robots around 0.1 mm. Testing is done at rated load, with a laser tracker or high-accuracy target, across at least 30 cycles. Simple refurbishers that only do cosmetic work commonly drift beyond 0.2 mm, which is far outside tolerance for electronics assembly, battery stacking or precise welding.

What a professional test report contains

The differences in the table are checkable in one conversation. Ask for the raw data, not a summary. Ask which instrument made the measurement. Ask whether the test ran at rated load. A refurbisher with laser calibration equipment answers those questions with files; a trader answers them with silence.

Why precision is the trust test for used robots

Precision is the intersection of every part of a robot's health: gearboxes, bearings, encoders, backlash, and alignment. A robot that meets its repeatability specification has passed the hardest exam a used machine can take. That is why our 12-point inspection includes precision verification with the machine under load, and why every robot ships with the report. The buyer who checks precision first rarely gets burned second.

Exceeding expectations: the report that closed the deal

The buyer who got burned now demands one thing before any purchase: the raw repeatability data from a laser measurement at rated load. On the first ZOTRADE machine they checked, the report showed 0.02 mm on the critical axis, better than the OEM spec, with the test procedure documented. They signed within the week. The machine has run the same precision job for two years without a rejected batch. A good report does not just prevent a bad purchase; it makes a good one feel certain.

When you ask for raw data, watch what comes back. A genuine report shows the measurement method, the number of cycles, the load condition and the instrument used, with values in millimetres. A suspicious report shows only a single number, or no mention of load, or a test run with empty air. Simple refurbishers often skip the rated-load condition entirely, which is exactly how a robot looks precise in a report and drifts on the line. The rated-load question is the fastest filter between a real test and a photograph of one.

Request the raw repeatability data and test procedure for any robot in our stock. Numbers first, then the invoice.

Need Industrial Robots or Urgent Replacement Spare Parts?

Our Guangzhou robotics facility stocks 10,000+ tested components with 24-48h global express dispatch and consultation-confirmed warranty.