Driving a screw takes about 0.4 seconds. Getting to the next screw usually takes longer — which is why cycle-time estimates built on screw count alone are consistently optimistic.
At a glanceDriving 0.37–0.41 s per screw on the arm platform; add for long threads and hard materials. · Travel Short moves are acceleration-limited, not speed-limited. Model accel and decel, not a constant velocity.
Most engineers arriving here go straight to one of these:
Build the estimate in three parts. First, driving: on our arm platform that is 0.37–0.41 s per screw, higher for long threads or hard materials, lower for short engagements. Second, travel: the time for the head to move from one screw to the next, which is total path length divided by the achievable axis speed — nominal speed is up to 800 mm/s per axis, but acceleration and deceleration on short moves mean the effective average is much lower, and the right way to model a short zig-zag is to treat each move as accel-and-decel rather than as a constant velocity run. Third, everything else: screw collection or feed presentation, fixture clamping and unclamping, and part load and unload.
The dominant error in most estimates is ignoring the third part. A station that drives twelve screws in five seconds can still have a thirty-second cycle because the part has to be loaded, clamped and unloaded around it. If the part is already on a conveyor inside an in-line station, that cost mostly disappears; if an operator is feeding a bench machine, it usually decides the cycle.
Tool changes and orientation changes have the same effect. Whenever the head must approach from a different angle, an articulated arm spends time reorienting, and a Cartesian gantry may need the fixture to reposition the part. Counting the number of distinct approach orientations on a part is often more informative than counting screws, because groups of screws sharing one orientation are cheap and each change of orientation is expensive.
Last updated 2026-11-20. Screw range, torque and cycle-time figures reflect the current published platform specifications.
The practical conclusions from the sections above.
0.37–0.41 s per screw on the arm platform; add for long threads and hard materials.
Short moves are acceleration-limited, not speed-limited. Model accel and decel, not a constant velocity.
Collecting or presenting the screw adds time per screw and is easy to overlook.
Load, clamp, unclamp and unload usually exceed all the driving time combined on a bench station.
Count distinct approach orientations, not screws — each change costs reorientation or a fixture move.
Screws sharing a plane can be driven simultaneously; this is the single largest cycle-time lever when screws come in groups.
Describe the part and the joint and we will tell you which configuration fits — or whether you do not need a machine at all.