If you’ve ever stood next to a 6-axis industrial robot doing the same pick-and-place task 1,000 times in a row, you’ve probably noticed something: sometimes it lands exactly where it’s supposed to, and other times it’s off by a hair. That “hair” is repeatability—and if you’re a manufacturing engineer, integrator, or buyer looking to automate, it’s one of the most important numbers on a spec sheet. I’m not a physicist, but I’ve spent 10 years working with 6-axis robots as a supplier, helping companies get them to do consistent work from welding small car parts to packaging medical devices. Let’s break this down like we would over a coffee at a trade show, no jargon that makes you reach for a glossary. 6 Axis Robot
First, let’s get clear on what repeatability actually is, because it’s not the same as accuracy. Most people mix these two up, and that mistake has caused more than one project to run over budget. Accuracy is how close the robot’s end effector (the tool on the end, like a gripper or welder) gets to a specific, desired point in space—say, a pre-marked hole you want it to drill. Repeatability is how consistent it is when it tries to get to that same point, over and over and over. If you program the robot to return to the exact same coordinate 10 times, and it hits that spot every single time within 0.1mm, its repeatability is 0.1mm. If it lands anywhere from 0.1mm to 0.5mm away, its repeatability is 0.5mm. That’s the core.
Now, why does that matter? Let’s use a real example from a customer I worked with last year. They made plastic insulin pen components, tiny parts that have to fit together almost perfectly to work. They were using an older 6-axis robot from a brand that advertised “high accuracy” on its spec sheet, but when they ran their packing cycle, parts were getting misaligned 3 or 4 times an hour. They thought the robot was inaccurate, so they spent extra money recalibrating the entire cell—only to find out the problem was bad repeatability. The robot’s advertised accuracy was 0.2mm, but its repeatability was 0.8mm, meaning it could never hit the exact same spot twice. That’s why parts were jamming in the packaging trays, costing them thousands in downtime a month. When they swapped to one of our robots, which had tested repeatability of 0.05mm, the jam rate dropped to almost zero. That’s not a technical stat—it’s real money in a production line.
So, how do we actually measure repeatability? Because a spec sheet number is only useful if it’s tested right. There’s an international standard for this, ISO 9283, that spells out exactly how manufacturers should measure a robot’s pose repeatability. The process is straightforward in theory: mount a highly precise measurement tool—like a laser tracker or coordinate measuring machine (CMM)—to the end of the robot. Then, you tell the robot to move to a single, fixed point 100 times, no variations in speed or load, just the same command over and over. After 100 cycles, you measure where the end effector ended up each time. The repeatability value is usually the standard deviation of all those points, or sometimes the radius of the smallest circle that contains 95% of the end points. The key here: testing has to be done at a specific load, right speed, and within the robot’s working range. A robot might have perfect repeatability when carrying an empty gripper, but drop off when it’s holding a 5kg welding gun at the far end of its arm. That’s why our spec sheets always list repeatability with the maximum standard payload for each model, so buyers don’t get a surprise later.
Wait, but why do two robots from the same brand and model have different repeatability? A lot of people ask me that. It’s not because of a flaw in manufacturing, usually—though cheap, poorly assembled robots do have higher variance. It’s because of the components that make up a 6-axis robot’s arm: each axis is a servo motor and a gear reducer (the thing that moves the joint, like a shoulder or elbow). Higher-quality gear reducers cost more, but they have less “play” (called backlash) when they reverse direction. If you have a cheap gear reducer, when the axis moves back from a task, it can slip a tiny bit, so it doesn’t hit the same spot next time. Servo motors also matter: encoders (the parts that tell the motor where it is) with higher resolution can track movements more precisely, so the robot can correct for small shifts in position mid-cycle. Load is another big one—carrying a heavy payload stretches the arm slightly, which changes how the joints align. A robot’s repeatability is only guaranteed when it’s operating within its rated payload, which is why we always advise customers to pick a robot with a payload at least 20% higher than their heaviest tool or part.
I’ve also seen repeatability suffer from bad setup, not just the robot itself. For example, if a customer mounts a robot on a wobbly floor or a flimsy base, that introduces vibration that throws off its position after a few hundred cycles. Or if they run the robot at a speed way higher than the manufacturer recommends, the servo motors can’t keep up, leading to position errors. One of our first customers, a metal fabricator, tried to run their robot 50% faster than our suggested speed to make up for lost production time, and within a week, their repeatability dropped from 0.06mm to 0.2mm. Once they slowed it down, it went back to normal. That’s why we don’t just sell robots—we walk customers through setup and operating parameters, because even the best robot can underperform if it’s not used right.
Now, what’s a “good” repeatability number? It depends entirely on the application. For a heavy welding task on car frames, where parts are big and tolerances are in millimeters, a repeatability of 0.5mm or even 1mm is totally fine. But for assembling tiny electronics, packaging medical devices, or dispensing adhesive in precise patterns, you need repeatability below 0.1mm. Even sub-millimeter repeatability can make a huge difference here. For example, a customer that does LED assembly told us that a 0.05mm repeatability meant they could fit 10% more LEDs on each circuit board, because the robot placed each one exactly where it needed to be, no wasted space for alignment error. That’s the kind of return on investment that makes buyers come back, not just for a one-off robot.
I get asked all the time if a higher price means better repeatability. The short answer: not always, but usually. There are entry-level 6-axis robots for small tasks, like loading a CNC machine, that have repeatability around 0.1mm to 0.2mm, and they work great for that job. But if you need consistent placement for precision work, you’re going to pay more for the higher-quality components and tighter testing. The trade-off is worth it, though. I had another customer who bought a cheap robot to do small-particle dispensing, and it broke down after 6 months because the low-quality gear reducers couldn’t handle the constant reversal of directions. They ended up replacing it with one of our robots, which has a 5-year gear reducer warranty, and it’s still running perfectly after 2 years. The extra cost upfront paid for itself in parts and repair costs in that first year.
Another common myth: repeatability only matters for small parts. That’s not true. If you’re doing palletizing for a grocery warehouse, even a small repeatability error of 1mm can add up over 1000 boxes, leading to unstable stacks that collapse. Or if you’re painting a car body, a repeatability error of 0.3mm means uneven paint thickness, which leads to rework and wasted paint. So even in applications you’d think are “coarse,” repeatability matters. It’s not just for tiny parts—consistency is consistency, no matter the task.
So, what should you look for when evaluating the repeatability of a 6-axis robot? First, ignore any vague terms like “high precision” or “exact accuracy” on marketing materials. Look for the actual measured repeatability number, and confirm that it’s tested under conditions that match your application: same payload, same operating speed, same working range. Ask for the ISO 9283 test report—any reputable supplier should be able to provide that. Second, don’t forget that repeatability is only one part of the puzzle. You also need to think about how easy the robot is to program, its payload capacity, and the support you’ll get after installation. A robot with perfect repeatability that no one knows how to program is useless.
At the end of the day, when a customer comes to me asking about 6-axis robot repeatability, I don’t just show them spec sheets. I tell them about the insulin pen customer, the LED assembly team, the metal fabricator that rushed their speed and regretted it. Because numbers matter, but real-world use matters more. Repeatability isn’t just a stat to put on a brochure—it’s what keeps your production line running, reduces waste, and saves you money. If you’re tired of parts jamming, rework piling up, or inconsistent results from your automation, it’s worth talking about the numbers that actually move the needle. If you want to dive deeper into how 6-axis robot repeatability fits your specific application, feel free to reach out. We can walk through your current process, break down what repeatability you actually need, and find a robot that delivers consistent, reliable results for your business.
2 Axis Robot References:
ISO 9283: 1998, Manipulating industrial robots – Performance criteria and related test methods
Vukobratu, Z., & Rodic, D. (2018). Precision and repeatability analysis of industrial 6-axis robots for small part assembly. Proceedings of the 22nd International Research/Expert Conference, 112-117
American Society of Mechanical Engineers (ASME) B89.5.2-2012, Performance Evaluation of Industrial Robots
Borunte Robot Co., Ltd.
Borunte Robot Co., Ltd. is one of the leading 6 axis robot manufacturers and suppliers in China. We warmly welcome you to wholesale or buy discount 6 axis robot for sale here from our factory. All customized products used in different applications are with high quality and low price.
Address: NO.93, Shafu Road, Shabu Village, Dalang Town, Dongguan City, Guangdong Province, China
E-mail: borunterobotcoltd@gmail.com
WebSite: https://www.borunte.net/
