Which Criterion Should Be Examined First When Selecting a Robot?
Robot selection usually begins with a brand or price comparison. The correct starting point, however, is the application itself: the weight of the part to be handled, the farthest point that has to be reached, the targeted cycle time and the required positional precision. Every comparison made before those four figures are clear is misleading.
In practice, the most common mistake is selecting the robot for today’s part only. If the parts likely to be produced in the same cell two years from now are left out of the calculation, the result is an investment that runs out of capacity within a short time. When selection criteria are defined, the baseline should not be the heaviest part and the farthest point of today, but the heaviest part and the farthest point of the next three years.
Why Must Gripper Weight Never Be Left Out of the Payload Calculation?
Payload is the total weight that can be hung on the robot’s wrist flange; it is not the weight of the part being carried. The distinction looks trivial, yet it is the source of the selection error encountered most often in the field.
The total load calculation has to include the weight of the gripper itself, the adapter plate connecting the gripper to the robot, the quick changer if one is used, camera and lighting brackets, sensors, and the weight of cables and hoses at the tool end. In a dual-gripper arrangement the robot briefly carries two parts at once, so the calculation must be made for that worst case. Once these items are added up, it becomes easy to see why a 10 kg class robot can fall short for a 5 kg part.
How Does Moment of Inertia Limit the Usable Carrying Capacity?
The payload figure in a catalogue is given for the ideal case, where the load sits close to the flange and on its centre. As the centre of gravity of the load moves away from the flange, the moment acting on the robot’s wrist axes grows and the real capacity drops.
For that reason a light part carried at the end of a long gripper can be more demanding than a heavier part carried with a short one. When the centre-of-gravity distance is ignored, the outcome is usually not a robot that refuses to run; it is a forced reduction in speed, vibration, and wrist wear that accelerates over time. Bringing the centre of gravity closer to the flange in the gripper design is often more economical than moving up to the next robot class.
Field note: Ordering a robot before the gripper design is finished is a risky sequence. The correct order is to define the part and gripper geometry, then calculate total load and moment, and only then select the robot class.
How Are Reach and the Working Envelope Verified?
Reach is not a single number. The catalogue value shows the maximum extension at the moment the robot stretches its arm fully, but at that point the robot is at its weakest, slowest and least precise. On top of that, the geometry of the gripper and the approach angle to the part pull the usable area noticeably below the figure on paper.
What actually has to be verified is not the distance but the entire working envelope: whether the robot touches the door, the fixture or the safety fence as it enters the machine, whether it can reach every target point at a reasonable joint angle, and whether it can avoid singularities. This check matters especially in CNC machine tending applications, because the opening of the machine door severely restricts the robot’s room to manoeuvre.
Are Repeatability and Accuracy the Same Thing?
No, and confusing the two creates the wrong expectations. Repeatability is the scatter between the points a robot actually reaches when it is sent to the same position again and again. Accuracy is how close the robot comes to the theoretical coordinate described in the program.
The repeatability of industrial robots is typically in the range of a few hundredths of a millimetre, and that is the value quoted in catalogues. Absolute accuracy is generally lower. For the vast majority of applications this is not a problem, because points are defined by teaching them to the robot. If the work involves trajectories generated directly from CAD data, or if part position comes from a camera, absolute accuracy enters the picture and the calibration requirement has to be planned from the start.
How Does the Cycle Time Target Shape Robot Selection?
Cycle time affects robot selection as directly as payload does, yet it is discussed far less. Between two robots doing the same job, the faster one can look like an unnecessary luxury; but if the targeted production quantity cannot be met, the entire investment misses its purpose.
The point to watch here is that cycle time is not made up of robot speed alone. Gripping and releasing time, the opening and closing time of the machine door, the acknowledgement time of safety equipment and part positioning time often make up more than half of the total cycle. Speeding up the robot delivers nothing on its own until those items are improved.
How Many Axes Are Needed: SCARA, 4-Axis or 6-Axis?
The number of axes is determined by how the part has to be oriented. More axes is not always better; unnecessary degrees of freedom add cost and programming complexity.
- SCARA: Very fast in planar, top-down assembly and placement work. If the part does not need to be tilted, this is the most efficient solution.
- 4-axis palletizing robot: Always keeps the load vertical. In case and sack palletizing it combines high payload with low cost.
- 6-axis robot: Mandatory whenever the part has to be oriented at an arbitrary angle. Machine tending, welding, dispensing and complex assembly fall into this class.
- 7-axis / additional-axis configurations: An advantage when the robot has to work around obstacles in a confined space.
How Do the IP Rating and Ambient Conditions Affect the Choice?
The environment the robot will work in is the heading skipped most often among selection criteria. Coolant mist in machining, dust in foundries, wash-down water and chemical cleaning agents in food production all act directly on the robot body.
A standard robot body is usually rated around IP54; applications with heavy liquid contact require a higher protection class for the wrist area in particular. In the same way, when ambient temperature rises above the robot’s permitted operating range, performance limiting kicks in. If these conditions are not taken into account during selection, the price of the investment comes back as a rising failure rate in the very first year.
Cobot or Industrial Robot: Which Data Drives the Decision?
This decision is not a matter of preference; it is the result of three data points: the required cycle speed, the required payload and the need for space shared with an operator.
Because a cobot can share the same area as a person, it reduces the need for safety fencing and fits into a small footprint. In return, its speed is limited by safety requirements. An industrial robot offers high speed and payload, but demands a defined safety zone around it. In low-volume, frequently changing work that runs next to an operator, cobot-based CNC automation systems stand out, while on high-volume lines with a fixed product an industrial robot almost always produces a lower unit cost.
When Is a Seventh Axis (Linear Rail) Actually Necessary?
A linear rail carries the robot beyond its own working envelope. The most common reason for using one is to let a single robot feed more than one machine or station.
There is a simple check for the rail decision: if the distance between stations fits inside the robot’s reach envelope, the rail is unnecessary and only adds cost and a maintenance item. If it does not fit, a rail is almost always more economical than buying a second robot. The extra travel time introduced by the rail must, however, be included in the cycle calculation; while the robot is moving, the machine that is not being served is sitting idle.
How Is a Robot Selection Validated Before Commissioning?
All of the criteria above can be checked by calculation, but calculation only verifies them one at a time; it does not show whether they are all satisfied simultaneously. A selection that looks fine on payload, sufficient on reach and on target for cycle time may still fail because it hits the safety fence in the real layout.
That is why the selection should be validated in three dimensions before the order is placed: the robot is run virtually with the real layout, the real gripper and the real part geometry, cycle time is measured and collision checking is carried out. A mistake that surfaces during simulation is corrected within a day; the same mistake found on site means rebuilding both the equipment and the schedule. Choosing the right robot is, in the end, not a catalogue comparison but the job of validating all the constraints of the application together.