Why a welding robot does not pay off on small batches
In small batches, it is not the arc that costs, it is the preparation: program, jig, tacking. Why conventional robots struggle to pay off, and what changes the game.
A welding robot welds faster, more consistently and for longer than a welder. Yet many workshops that bought one see it run far less than planned. The reason fits in one sentence: in small batches, arc time is only a small part of the work.
Breaking down the time for a welded part
Take a typical small-batch part: a steel bracket made of six laser-cut components, about one metre of weld. On a part of this type that we studied, we estimate the working time at about 21 minutes per part, of which only 3 minutes are arc time. The rest:
- reading the drawing and checking the components;
- placing the parts in the jig, or positioning them square;
- tacking the assembly to make it rigid;
- turning the part over, inspecting, correcting.
These figures are an internal estimate, which we will confirm by timing on the cell. But every workshop manager knows the order of magnitude: preparation weighs far more than the welding itself.
What a conventional robot automates, and what it does not
A conventional welding robot, or a cobot, automates exactly those 3 minutes of arc. For the rest, it even adds work:
The program
Every new part number has to be programmed: teaching points on the pendant, or offline programming from CAD. Depending on the complexity of the part and the programmer’s experience, this takes from a few tens of minutes to several hours. On a batch of 500 parts, it is negligible. On a batch of 10, it often takes longer than welding by hand.
The jig
For the robot to find the joint in the right place, the part has to be positioned repeatably. That means a jig or clamping fixture, to be designed, built, set up and stored. Again, the cost is spread over the batch: it becomes prohibitive when you change part number every week.
Tacking
In most installations, the components are placed and tacked by hand before going to the robot. The qualified welder you wanted to free up stays tied up.
The maths that does not add up
To sum up for a short batch:
| Task | Manual welding | Conventional robot |
|---|---|---|
| Programming | none | per part number |
| Jig | sometimes | almost always |
| Placing and tacking | welder | welder |
| Arc | welder | robot |
The robot only saves time on the last line. If the first three dominate, the gain disappears. That is why welding robots first conquered the automotive industry and large series, and why “high-mix, low-volume” workshops (many part numbers, small quantities) remain mostly manual.
To put a figure on what preparation costs in your workshop, use our ROI calculator.
What changes the game
To make a robot pay off on small batches, you need to automate not the arc, but the preparation. Three technological building blocks make this possible today:
- Automatic programming from CAD. If the software can read the 3D model, find the welds and compute collision-free paths, the program no longer costs anything per part number.
- Positioning without jigs. Magnetic tables, a second robot arm holding the part, tab-and-slot components: we describe these solutions in welding without jigs.
- Automatic tacking. The robot tacks the assembly itself, in an order calculated to limit distortion. See why tack welding determines quality.
Putting these three building blocks together is what makes an autonomous welding cell: changing part number means changing file.
Should you give up on cobots?
No. For a workshop that often welds the same parts in medium to long runs, a cobot is an excellent investment, simple and affordable. The trap is buying one for very varied small batches, hoping it will free up welders: it then shifts the work more than it removes it. Our cobot or autonomous cell comparison helps you choose.
Key takeaways
In small batches, the question is not “does the robot weld faster?” but “who prepares the part?”. As long as programming, jigging and tacking remain manual, a robot cannot pay off on lots of 10 or 50 parts. Automating the preparation is the condition for automation to finally benefit small workshops.