Buying an industrial robot is one of the most exciting and most reversible mistakes a factory can make. The robot that reaches, the payload that suits the gripper, and the cycle time that actually fits the line all live inside a few pieces of anatomy that every automation engineer should read like a map. This tutorial explains the structure of an industrial robot, the kinematics that give it its shape, the wrist design that allows orientation, and the workspace math you need before you draft the purchase order.
🤖 Think of the robot as a set of joints and links arranged to trade reach for stiffness and agility. The layout of those joints, called the kinematic structure, decides what the machine can and cannot do, long before any programming begins.
1. Serial vs Parallel Kinematics: The Two Big Families
Most factory robots are serial kinematic machines: a chain of links connected by rotary or prismatic joints, one after the other from the base to the flange. The articulated 6-axis robot, with its base, shoulder, elbow, and three wrist axes, is the celebrity of this family because it combines long reach with the ability to orient the tool in any direction. Serial arms are simple, reach far, and can carry moderate payloads, but they pay for that reach with flexibility: the further the load sits from the base, the more the arm bends under its own weight.
Parallel kinematic machines, such as delta robots, hold the tool with several closed-loop chains of links. The closed loops make them extremely stiff and fast, ideal for light payloads and high cycle rates, which is why delta robots dominate small-part pick-and-place lines. Their penalty is a small, dome-shaped workspace and limited orientation. The choice between serial and parallel is a trade of reach and dexterity against speed and stiffness.
2. Joint Configuration Types You Will Meet
- Articulated (6-axis or 4-axis): general-purpose, welding, assembly, material handling.
- SCARA: horizontal assembly champion, rigid vertically, fast in the plane.
- Delta: high-speed light pick-and-place and packaging.
- Cartesian gantry: large workspace, heavy payload, three linear axes.
- Collaborative cobots: human-safe joint torque limits, small shops and flexible cells.
Each configuration exists because some real application needed exactly that mix of reach, speed, stiffness, and cost. Match the configuration to the task envelope, not to the sales brochure.
3. The Six-Axis Wrist: Why Three Axes at the End
The wrist is where the robot earns its title. Three additional axes, called roll, pitch, and yaw (or axis 4, 5, and 6 on a six-axis arm), let the flange orient the tool in any direction, which is exactly what a welding torch, a spray gun, or a six-sided machining head needs. The classic wrist is a roll-pitch-roll arrangement whose axes all intersect at one point, the wrist center point. That intersection simplifies the inverse kinematics enormously, which is why the design appears on nearly every articulated robot on the market.
Wrist design is a compromise. Bigger wrist motors give more dexterous payload and faster reorientation but add weight that the arm must carry. Wrist interference is the subtle killer: a wrist that cannot reach the required orientation because a link hits the tool cable. When you validate a robot, always check the wrist orientation limits, not just the arm reach, because many applications die in the wrist plane.
4. Reach, Workspace, and the Reach Envelope Trap
The workspace is the volume the flange can reach, and it is never a simple sphere. For a six-axis arm it is a complex volume with a spherical-zone shape around the base, often drawn as a cross-section in the catalogue. The max reach number is a headline; the useful reach at the right orientation, over the work height, is what your cell cares about. A robot whose max radius is 1200 mm may hold useful payload comfortably at 900 mm and barely at the edge.
Begin the layout by drawing the pick and place positions as points in space with the required orientation, then fit the smallest arm whose reach and wrist envelope cover both points with margin. Compute reach from the base to the tool center point, not to the flange, because the gripper length eats reach quickly. Include the conveyor height, the fixture height, and the stroke of the end effector in the same sketch, and the workspace becomes obvious long before the quotation.
5. Payload, Moment, and the Force at the Flange
The payload rating tells you what the flange can carry at full reach, but real payload falls as the tool gets longer or heavier. Two secondary numbers matter more: the allowable moment at the flange and the allowable inertia around the wrist axis. A long gripper with a small part can exceed the moment limit even when the weight is fine, because the load arm multiplies the force. Compute the moment of the tool plus the part about the flange center, and compare it with the datasheet values for the wrist at the working speed.
Dynamic load matters as much as static. During a quick move, acceleration multiplies the effective load several times, so a robot rated 10 kg at 1 g acceleration may see an effective 25 kg at the tool tip during a hurry-up pick. Slow down the acceleration or shorten the reach instead of buying a bigger robot that then moves too slowly for the cycle. Every robot sizing is really a conversation between reach, payload, acceleration, and cycle time.
6. Speed, Accuracy, and Repeatability: Know the Difference
Repeatability is the robots ability to return to the same taught point and is usually the headline number, often plus or minus 0.05 mm on a good six-axis arm. Accuracy is the robots ability to reach a point commanded from coordinates in space, and it is always worse, because it accumulates calibration and temperature errors. For most pick-and-place work, teach the points and repeatability rules everything. For offline-programmed tasks or vision-guided bins, accuracy begins to matter, and a calibration routine or a vision system compensates for what the robot cannot guarantee.
Cycle time is the third number that decides feasibility. The datasheet gives point-to-point times for a reference move, but your real cycle adds the acceleration, the wrist reorientation, the gripper open and close, and the safety speed near people. Model the motion as a trapezoidal velocity profile: accelerate, cruise, decelerate, and you will predict the true cycle within ten percent, which is far better than trusting the catalogue curve.
7. The Pick-and-Place Cell Case Study
Design a cell that takes a 300 g part from a conveyor, turns it 90 degrees, and places it in a tray, twelve times per minute. The stations are 150 mm apart and 200 mm apart vertically. Reach from the base to the tool is under 800 mm at the farthest point, and the wrist must roll 90 degrees mid-flight. A SCARA or a small six-axis robot both fit the workspace; the SCARA wins if the orientation is fixed, the six-axis wins if the tool must approach from odd angles.
With a total move of roughly 250 mm and a cycle budget of 5 seconds including the turn, even a modest robot achieves it easily. The real constraint appears in the gripper: the 300 g part extended 180 mm from the flange creates a moment that must stay under the wrist rating. A short, lightweight gripper with a vacuum cup beats a long ugly gripper every time. This is the classic case where the robot is over-specified and the tooling is under-thought, and the fix is layout, not hardware.
8. Grippers and End-of-Arm Tooling
The end effector is where the automation actually meets the product. Vacuum grippers cup the part with suction, ideal for flat, clean, nonporous surfaces; finger grippers grip and reorient, at the cost of weight; magnetic grippers lift ferrous parts quickly but lose the nonmagnetic ones. Sizing the gripper means checking the grip force against the weight times acceleration plus a safety factor of three or more, because a dropped part at high cycle rate is a production loss and a safety event.
Tool-change plates argue for themselves when the cell runs multiple products. A robot that swaps a gripper in seconds handles a broader product mix with one arm, but every gram of the quick-change plate reduces the payload for the real tool. Balance the flexibility gain against the payload loss before you add the plate.
9. Safety, Programming, and Commissioning
A robot cell is a safety system as much as a machine. Guards, light curtains, and safety-rated stops protect humans, and the risk assessment, not the budget, decides which ones you need. For collaborative robots, the speed and force limits let a human share the space, but the application still requires a documented risk review, because a cobot with a sharp gripper is no longer harmless.
Commissioning follows a fixed rhythm: jog the axes slowly in manual, verify the home positions, teach the picks and places, run the motion at reduced speed, then raise the speed while watching the cycle timer. Log the final points and the controller backup on day one, because the robot that works reliably is the one whose configuration is documented. Programming the robot with vision and a PLC handshake adds the last layer, and a well-defined interface between the robot controller and the line PLC is worth more than any library of fancy motion.
10. Maintenance That Keeps Kinematics Honest
Robots are maintenance-hungry machines that reward discipline. Check the gearboxes for backlash growth with a dial indicator at the wrist, watch the joint temperatures for a quietly failing reducer, and inspect the cable harness where it flexes, because a chafing cable at the elbow fails without warning. Backup the controller and the calibration files on every visit. A calibration that drifts by a few tenths of a millimeter is invisible on one part and catastrophic on a stack of two hundred.
Conclusion
Industrial robot anatomy, read correctly, turns a big capital purchase into a layout problem you can solve on paper. Choose serial or parallel kinematics for reach and speed, respect the wrist orientation limits, compute the real reach to the tool center, check the moment and dynamic load at the flange, and keep repeatability, accuracy, and cycle time as three honest numbers. The pick-and-place case study resolved itself in a page of layout math, and the moment check decided the gripper. Whether you buy a SCARA, a delta, or a six-axis arm, the anatomy is the same story: the joints define the reach, the wrist defines the dexterity, and the tooling defines the cell. Apply this structure to the next robot you quote, and the machine will fit the task it was bought for.
Rule to remember: robots are sold by reach and payload, but they are justified by wrist dexterity and cycle time. Measure both before you sign.
