Preparing the published article.
Cycle-Time Claims: Define the Test Conditions First
How to write measurement conditions, control the test batch and record results before accepting a quoted robot cycle time.
Write the Measurement Conditions Before the Demonstration
A cycle-time figure is only meaningful with its conditions attached, so write a one-page measurement sheet before any demonstration and have both parties sign it. Record the exact part number and revision, its mass, its centre of gravity, and the tolerance band of the incoming parts. Record the start event and the end event of the cycle: is the clock started by a part-present signal at the infeed and stopped when the part is clamped in the fixture, or when a good-part signal is returned downstream? Those two definitions can differ by several seconds, and neither is wrong, but mixing them across suppliers makes the comparison meaningless. Also record which signal is authoritative for the start event, a PLC bit, a sensor input or a manual push, because a manual start can quietly add operator reaction time to every cycle.
Also fix what counts as part of the cycle. Ask explicitly whether the quoted time includes gripper open and close, tool change, vision acquisition and processing, machine tending, part release, and return to the home position. Ask whether it includes a re-grip or a second approach when the first pick fails. If the answer is "typical" or "optimistic case", ask for the slow case over a defined batch and for the number of consecutive cycles that were measured to produce it. Ask as well whether the figure describes a single part or a family, and if a family, which member was measured and which member is slowest.
Control the Rig, the Operator and the Batch
Decide where the measurement happens: supplier floor, your floor, or a third-party facility. Record who programs the motion, who sets the speed and acceleration limits, and whether safety-rated speed limits are active during the test. If the robot runs faster only because guarding and safety functions were disabled for the demonstration, note in the record that the figure does not represent production conditions and cannot be used for capacity planning or for a line balance calculation.
Control the material as well as the machine. Use parts from your own production rather than demonstration parts, and state the batch size, the number of operators involved, and whether an operator loads the infeed by hand. Log each cycle with a timestamp so you can inspect the distribution rather than the headline. Ask for the minimum, median, maximum and the count of cycles that exceeded an agreed threshold. An average without a spread hides the stoppages that decide whether the cell meets the shift target. Note also how the infeed is replenished during the run, because a test that pauses while someone refills a magazine measures a cell that will never exist in production.
Record Results, Exclusions and Remedies
After the test, write down the conditions as they actually were, not as they were planned. Note ambient temperature, air pressure at the gripper, part temperature, and any manual intervention such as clearing a jam or re-seating a part. Record every fault, its duration, its cause where known, and who cleared it. If the measured time misses the target, agree in writing what will change, who pays for the change, and how the retest will be run and witnessed.
Annex the results sheet to the purchase order so the number becomes a contractual reference rather than a memory of the visit. Add a clause describing what happens when production conditions differ from the test conditions: which party investigates, within how many working days, and what evidence is collected. Finally, plan a verification run after installation on live parts, and define in advance how many shifts of data you will collect before declaring the cycle time accepted. State the threshold in the same units as the original claim, so nobody later argues about whether the target was per part, per cycle or per shift.
Photo credits: the images in this article are licensed reference photographs of the relevant robot class, credited with author and licence in each caption; they are not photographs of the specific equipment discussed. Licence and author details are recorded with each image.