A robot can follow the same path around a metal disc every time and still produce inconsistent joints if the parts do not present the same gap and position to the torch. Before changing the robot program, separate three questions: does the joint start in a repeatable condition, can the torch reach it as planned, and what happens to the assembly over the chosen welding sequence? A representative trial should record each of those conditions alongside the finished joint.
Start with the joint, not the robot path
Imagine two discs that look alike in a drawing but sit differently against their mating hubs. The robot may repeat its taught motion faithfully while the gap under the torch changes. A wider or wandering gap can change how the weld forms, so asking the programmer to “correct the circle” may send the team after the wrong variable. General welding guidance puts fit-up, excessive gaps, clamping and fixturing among the first things to check when welds misbehave. The relevant question for a disc family is whether its locating faces and the joint edge arrive in the same relationship each time.
For a trial, mark a datum on the disc and its mating part. Record the joint gap and seating at several positions before any arc is struck, then remove and reload a representative part. If the readings change after reload, resolve part presentation or fixture location before treating the programmed torch path as the sole cause. That observation identifies which variable to address before revising the taught path.
The filmed example shows a robot torch striking an arc repeatedly over a disc-shaped metal part at a fixed station and moving along the disc surface. It gives a useful view of the torch-to-part relationship. The buying decision must then use the actual joint drawing and sample parts.
The torch needs a usable position throughout the intended weld
Once the part repeats in its fixture, check the torch position at the most awkward points of the intended joint. A route that clears the part at one angle may put the torch, cable or nozzle too close to a rim, hub or clamp elsewhere. Map approach, welding and retreat positions with the real tool fitted. If a clamp blocks a required orientation, changing a few taught points may be less useful than changing where the part is held or how it is presented.
This check also separates a drawing’s ideal circle from the actual workpiece. If the joint location changes with part or fixture variation, define the allowable variation and the means of locating it before promising a single taught program for the whole family. Set that range with the real parts and fixture.
Treat welding order as a trial variable
Weld sequence and restraint can influence distortion in welded assemblies. Published welding guidance discusses tack welding, joint gap and sequence as ways to control movement while welding. The right sequence for this disc joint must be decided by a representative trial. For the proposed disc joint, the production team should compare candidate sequences against its own drawing and acceptance criteria.
A practical trial record pairs each sample with its initial seating, clamp condition, weld order, and measured state after cooling. If two samples start from different joint gaps, their finished geometry cannot fairly decide which robot path is better. If they start alike but differ after a sequence change, that result points the next trial toward welding order, restraint or heat input. The acceptance method should be agreed before production programming, whether the priority is joint continuity, dimensional condition or both.
What to bring to the first engineering discussion
Bring a joint drawing, representative discs and mating parts, the allowed fit-up range, photographs of the present holding method, and the actual acceptance requirement. EVST can use those inputs to discuss the workholding concept, torch access and a focused trial plan. The first useful answer is which variable to test first and what observation would change the cell design.