Jewelry Welding on the Production Side: The Answer Has to Survive the Next Pair of Hands
One piece leaves your bench and comes back with what was done to it written down — because the person running the batch next week may not be the one who ran it today.
Micro welding for jewelry manufacturing — hollow and thin-wall components, chain and findings assembly, settings joined as a production step.
Most of this page is about the part of the answer that has to be written down rather than remembered. Not that a joint can be made once. Whether the same joint can be made the same way by someone who was not standing there the first time.
Production jewelry: the parts we weld, and the metals they come in
A parts list and a materials list — and one line about what is not on either.
Hollow and thin-wall components. Chain and findings assembly. Settings joined as a production step. These arrive in runs rather than one at a time, which is what makes them different from a piece brought in for repair.
Gold, including karat golds
Same-composition, same-colour joints are the routine case — eighteen karat to eighteen karat.
Silver
Routine in the sense that it turns up constantly on a production floor, and not routine in the sense that it goes through a test weld before we quote it. The reason is in the limits section below.
Platinum
Platinum to platinum, sterling to sterling — the joint stays inside one composition and one colour.
Rolled gold and gold plate are not in the direct-weld group at all. The base underneath is doing the structural work and the layer on top is doing the appearance, and those two do not respond to the same energy. Those pieces go through a test weld first, for the same reason zinc-bearing solder does.
Jewelry welding is clamped between assembly and the finishing bench
Assembly is already behind it. Three finishing operations are still ahead of it.
A production piece is cast, cleaned up and filed, assembled, welded, then polished, surface-treated and stone-set before it is checked and shipped. Jewelry welding sits in the middle of that, and it is clamped from both ends.
Behind it
By the time a joint reaches us the piece has already been assembled — links closed, findings attached, seats seated. Assembly is what fixes the geometry the fixture has to grip, and a piece that cannot be held rigid cannot be welded the same way twice.
The weld
What is behind it decides whether the piece can be held. What is ahead of it decides whether a mark can be absorbed.
Ahead of it
Polishing, plating or oxidising, and stone setting all come after the weld, and each of them either takes the trace of a weld out of sight or puts it permanently on display. That is the whole reason the order of operations matters more here than the setting on the machine.
There is one case where this chain does not apply. If stones were already set before the joint reaches us, the sequence is broken and the piece stops being a production step — it becomes a piece judged on its own. That case is the third entry in the limits section below.
What the weld sits next to here is an operation that already happened
Not something sitting beside the joint — something already done to the piece.
On other benches the thing that cannot take the heat is sitting next to the joint while you work. Here it is usually finished and gone: the neighbour is the operation before this one.
Stones already set
On the production side a piece can reach the weld with stones already in it. What that changes is not a matter of degree, and it is handled separately below.
Faces already polished or plated
A polished flank and a plated surface both carry the record of what happened to them. A mark left near the joint is not discovered at the weld — it is discovered at the polishing bench afterwards, and by then the piece is finished work rather than work in progress.
Enamel and oxidised finishes
These are not surfaces that can be brought back to the same state by redoing them. Whatever they looked like was produced once, under conditions that are not repeated for a single piece.
Assemblies that still have to move
Chain links, folding clasps, hinged parts — by the time the joint reaches us these have already been put together, and they were put together to a fit. Anything that changes that fit changes a piece that was finished on the previous bench, not a blank.
Three materials we test-weld before we quote them
Three different reasons, and the same answer at the end of all three.
Three things on a production floor go through a test weld before we put a price on the work. They are not ranked, and they do not have a common cause. What they have in common is only where they end up.
Sterling silver
At our wavelength silver starts out absorbing very little of what arrives, so the first part of the job is getting the energy in at all. The failure that follows is the awkward kind: nothing, nothing, nothing, and then straight through — with very little between those two states. So silver gets a test weld and a written parameter set, and the set belongs to that metal from that supply. A different melt of the same alloy does not inherit it.
Zinc-bearing solder
Here the zinc is in the filler rather than on the surface, which changes where the trouble starts but not that it starts. Beyond the test weld, this one often needs the joint itself reshaped so the vapour has somewhere to go. We do not promise it comes right the first time.
Stones already set
The same energy gives completely different outcomes depending on the stone: filled, treated, organic, water-bearing. And the treatment history usually is not visible — a stone that has been through something rarely looks like it has.
All three end in the same place: one piece, welded and judged, before the batch moves. For the stones in particular, we do not publish a safe-distance rule, because there is not one that survives contact with a treated stone. A published rule would hold on most pieces and fail on the one nobody had flagged — which is the piece you would least have prepared for. What we give instead is the piece back with what was found written down, and that is something you can hand to the next person. Manufacturing-side work only. → start a free weld sample evaluation
More than one operator, and the weld parameters have to travel
Consistency that lives in one person's hands is not consistency yet.
Jewelry welding on a production run is not done by one person from start to finish. It passes between operators, across shifts, and over weeks. That is the condition this page is really about.
What it means in practice is that a good result is not enough on its own. If the only place the settings exist is in the hands of whoever got it right that morning, then the run holds together for exactly as long as that person is at that bench. The next person starts from a blank page and gets a slightly different joint, and the difference shows up as a batch that is not quite uniform rather than as a failure anyone can point to.
So what we hand back is not only a welded piece. It is the joint it was established on, and the settings that produced it, written out — because a setting that cannot be reproduced was never a parameter.
That is also why a parameter set is tied to the metal it was established on. A different melt of the same alloy is a new starting point, not an inherited one, and treating it as inherited is how a run drifts without anyone deciding to change anything.
Two of the four micro welders, for two different reasons
Two of the four, for two different reasons — and the choice itself is settled by a sample, not by this page.
If you can't see it, you can't place it — and where a bench has multiple operators and long sessions, that condition points at the High Value Type. The reason on production jewelry is its own: the same batch passes through more than one pair of hands, and seeing the joint properly is what makes it possible to hand the same judgement to the next person. A joint that one operator can place by feel is not a joint the run can rely on.
For thin-walled or heat-sensitive parts, low heat input and line cadence point at the QCW. The reason here is also specific: a hollow component that collapses is not something you straighten out afterwards. There is no recovering the form once it has gone, so the heat has to stay where it was put the first time.
Which one belongs on your bench is a question about the joints you actually run, and it gets settled by welding one of them rather than by reading a page. Where that decision has already been made on your floor, put it on a quote.
One piece in the post, and it speaks for the whole jewelry run
One or two real pieces, welded and written up, before the run commits.
You do not have to send the run. Send one or two pieces that are representative of it — the joint you actually make, in the metal you actually use — and we weld them and write down what it took.
What comes back is the answer to the question the batch is about to ask, in a form the next person on that bench can use. Whether the pieces are returned or kept is yours to decide, and the arrangements for sending and returning are settled in our first reply.