
A weldable housing is two parts, and the weld is where the design succeeds or fails
Weldable stainless steel sensor housings exist because some sensing arrangements cannot be made from a single drawn shell. A sensor that must reach into a vessel wall, a probe that has to be inserted to a depth, a housing that has to carry a flange and also be small, or a design where the cavity has to be hermetic and the geometry would need a two-sided draw – all of these end up as a welded assembly. And once a part is welded, the weld becomes the design’s weakest point, its only permanent joint, and the source of almost every dimensional surprise in the process.
We form the components, then laser or TIG weld them, dress the weld and leak test, then finish. Hydraulic drawing presses and CNC lathes for sealing faces, first articles and batch samples checked on the CMM, with turned or milled features to +/-0.01 mm and formed details to +/-0.03 mm. Standard stock is carbon steel in SPCC and DC01 grades, 304 and 316 stainless steel, aluminium and brass, with grade and temper confirmed in writing before tooling. Your drawing sets the part, DFM notes come back before any tooling is cut, and the same drawing is used for every repeat order. This page is about what welding changes in a sensor housing, starting with the fact that it changes the tolerances.
Welding distortion, and why the drawing has to accommodate it
A weld shrinks as it solidifies, and it pulls everything attached to it toward itself. On a small, thin-walled, machined part this produces distortion that is visible in dimensions that were held to a tenth of a millimetre before welding. Three consequences, and they shape the drawing:
- Tolerance on welded dimensions has to be looser than tolerance on machined ones. The +/-0.01 mm band applies to features that are machined after welding, or that are not affected by the weld’s shrinkage. A feature machined before welding and then distorted by it cannot hold that band. Where a tight feature crosses a weld, the drawing should either move the weld away from it or flag the feature differently.
- Part geometry affects distortion more than anything else. A round part distorts far less than a square one – a circular section has no corners to pull against, so welding it is close to isothermal. A long, thin, asymmetric assembly distorts the most. If distortion is going to be a problem, changing the assembly’s shape is a more effective fix than tightening the weld procedure.
- Weld sequence and fixturing are design inputs. A housing that can be welded in a fixture that supports it symmetrically behaves very differently from one that must be welded free. This is a manufacturability question and it is worth asking before the tool is cut.
Which weld, and why
Two processes appear on this part, and the choice is usually made for reasons beyond the weld itself:
| TIG (GTAW) | The conventional choice for stainless steel, giving a clean, controllable, well-behaved weld with full penetration and no spatter. Slower than laser, and it is a manual or semi-automatic process whose quality depends heavily on the welder. The advantage is that almost any stainless steel thickness and any joint geometry can be welded. |
|---|---|
| Laser welding | Faster, more repeatable, and produces a narrow heat-affected zone – which matters because a narrower HAZ means less distortion and less grain growth in the corrosion-critical area. It requires good joint fit-up and is less forgiving of gaps than TIG. Where a high volume justifies the equipment, it is the more consistent process. |
For a stainless steel housing the metallurgical requirement behind both is the same: the weld and its heat-affected zone must have corrosion resistance comparable to the parent metal. This is why filler selection matters and why a weld that looks clean can still fail – an unsuitable filler with lower chromium or nickel content will rust first at the weld, which is a counterintuitive and expensive failure mode.
Joint design: the four things that decide weld quality
On a drawn housing, the joint geometry is set by the fact that the components are thin shells. Four things control whether the weld leaks, and all four are drawing decisions:
- Joint fit-up. The gap between the two parts before welding. A tight, controlled gap is the single biggest factor in weld consistency, and it is controlled by the tolerance on the mating features – so the drawing’s tolerance on those features is really a weld-quality specification. A nominal-fit joint on a small stainless steel housing will produce variable penetration from unit to unit.
- Edge condition. A sheared or drawn edge carries work hardening and, in stainless steel, an oxide and heat tint. Welding a dirty or oxidised edge risks both porosity and inclusions, and inclusions are where corrosion and fatigue cracks start.
- Wall thickness at the joint. Where a thin shell meets a thicker section, the weld has to fill a step, and the step is where porosity appears. Where the assembly has two similar thin walls, the joint is far simpler. Designing the two components to similar wall thicknesses is a real cost saving on a welded housing.
- Corrosion at the crevice. A welded stainless steel joint in a wet environment has a crevice on one side and a bead on the other, and crevice corrosion can start there even in 316. Where the assembly will sit in a chloride environment, the design should avoid presenting a crevice or specify a passivation step after welding.
Leak testing as a process step, not a final inspection
The route here is weld, dress, then leak test. That sequence is deliberate and the ordering matters. Testing a part before dressing finds leaks in the weld’s rough surface – which is where a surface-breaking pore will show. Testing after dressing finds leaks in the finished, welded surface, which is the state the part ships in. Where a part has both a welded joint and a machined sealing face, testing at the end validates both in one operation, and a failure there can be attributed by examining which interface the leak came from.
Three practical points on leak testing a sensor housing specifically:
- Test pressure versus service pressure. A sensor housing may be tested well above its working pressure, and over-testing a thin drawn shell can find nothing useful while sometimes deforming a wall. The test pressure should be specified and it should relate to the design pressure.
- Helium versus pressure-decay versus bubble. Helium is sensitive and slow. Pressure decay is fast and good for gross leaks but can be fooled by trapped gas. Bubble testing is the cheapest and appropriate for a low-pressure or large-volume part. The right method follows from the required sensitivity, and that is a stated requirement rather than a shop decision.
- 100% test or sampled. On a hermetic sensor housing this is a design decision with a cost attached. For a housing where a leak causes a safety or measurement problem, full testing is usually non-negotiable.
Tolerances after welding: what is tight and what is not
The +/-0.01 mm on turned or milled features and +/-0.03 mm on formed details structure tells you exactly how the drawing should be written, and the rule follows from welding’s behaviour:
- Features machined after welding can hold the tight band, provided they are far enough from the weld that the local distortion has settled. This is the normal approach on a welded housing: weld the shells, machine the sealing face and the bore afterwards.
- Features machined before welding will move. If a bore has to be machined before welding for access reasons, its tolerance needs to accommodate the distortion of the welding operation, and the drawing should flag that rather than expecting the pre-weld band to hold.
- Overall dimensions are limited by the weld, not by the machining. The formed details’ band of +/-0.03 mm is the honest number for a dimension that spans a joint.
- Capability is checked on the first article before the run continues – and on a welded part this check is doing more work than usual, because it is validating the weld’s repeatability as much as the machining’s.
Finishing a welded stainless steel housing
Zinc, nickel or chrome plating, anodising, passivation or e-coat are available, with alternatives quoted alongside and samples approved in writing. For a welded stainless steel part the finishing logic is narrower and the order matters:
- Pickling and passivation of the weld. A stainless steel weld carries heat tint and an oxide layer that is less corrosion-resistant than the surrounding metal. Pickling removes it; passivation restores the surface. A welded stainless steel housing that will see moisture needs both, and both need to be written on the drawing.
- Plating a welded stainless steel part is possible but awkward. The weld bead and the surrounding sheet are surfaces of different geometry, and plating coverage on a bead is a local thinning. For a corrosion-critical duty, solid stainless steel with proper passivation is usually the better answer.
- Masking. Any sealing face or thread that must not be coated. On a welded part the masking has to survive the welding, so it is applied after welding and before finishing – which fixes the route.
- Electropolishing is worth considering on a welded stainless steel part where appearance and cleanliness both matter, because it removes the weld’s heat tint more evenly than mechanical polishing and leaves a smoother surface.
Tooling, samples and repeat orders
Tooling is built and held under your part number so tooling lead time stays under our control, and first-article timing is measured from the tooling start date – which on a welded part is a good deal earlier than the part’s completion date, because the assembly tooling has to be ready before the first weld. Prototype and pre-production samples are provided against the drawing, the part is proven before it is made in volume, and mass production is gated on that approval.
Seven things that make a welded sensor housing repeatable:
- Which features are machined before welding and which after. This determines where the tight tolerances can honestly apply.
- The joint fit-up tolerance, stated – because it is the main variable in weld quality.
- Weld process and filler, named on the drawing rather than left to the shop.
- Pickling and passivation after welding, if the housing will see moisture.
- Leak test method, pressure and whether it is 100%, as a specified requirement.
- Assembly geometry – whether the joint is a circular section, which distorts least, and where the joints sit relative to tight features.
- Filler material specified, because an unsuitable filler in stainless steel corrodes at the weld before it corrodes anywhere else.
And the design question worth asking before all of that: can the housing be one drawn piece instead of a welded assembly. It is not always possible – a two-sided draw, a deep cavity and a small package can rule it out. But where it is possible, a single drawn shell removes an entire class of cost, distortion and leak-path problems, and it is the cheapest insurance available in this design.
Sensor Housings is the parent range for parts like this one; the overview page gathers the tooling logic, finishing options and tolerance notes that apply across the family.
Specifications
| Item | Weldable Stainless Steel Sensor Housing |
|---|---|
| Brand | AUNOV |
| Designing | Your drawing sets the part – DFM notes are returned before any tooling is cut; the same drawing is used for every repeat order. |
| Tooling | Tooling is built and held under your part number, so tooling lead time stays under our control, and first-article timing is measured from the tooling start date. |
| Samples | Prototype and pre-production samples are provided against the drawing, so the part is proven before it is made in volume, and mass production is gated on that approval. |
| Materials | Standard stock: carbon steel (SPCC / DC01), 304 / 316 stainless steel, aluminium, brass; grade and temper are confirmed in writing before tooling. |
| Surface finish | Zinc, nickel or chrome plating, anodising, passivation or e-coat; alternatives quoted with the quotation, samples approved in writing. |
| Processing | Laser or TIG welding of formed components, dressing and leak testing, then finishing. |
| Equipment | Hydraulic drawing presses and CNC lathes for sealing faces; first articles and batch samples are checked on the CMM. |
| Tolerance | +/-0.01 mm on turned or milled features, +/-0.03 mm on formed details; capability is checked on the first article before the run continues. |
| Delivery terms | FOB, with DDP offered for EU and US addresses, T/T with deposit and L/C for larger runs; price breaks tie to annual volume. |
Frequently Asked Questions
What makes a stainless steel sensor housing weldable?
The grade is picked for both forming and joining – a low-carbon grade draws cleanly and welds without cracking, so the shell and any welded feature come from one material plan. The welding is located off the drawn datum so the joint stays related to the rest of the part.
How is the weld joint located and sealed on a sensor housing?
The joint is a formed or machined feature positioned off the shell datum, and the seal runs on the prepared face or a fitted insert. Because a sensor housing often sits in a pressure or ingress boundary, the weld is verified, not just visually passed.
Which stainless steel grade welds best for a sensor housing?
A low-carbon 304 or 316 grade is the usual choice, because it draws, then welds without sensitization in the heat-affected zone. The grade is chosen for the environment and the weld together, and certificates accompany the first article.
What is the minimum order quantity for weldable sensor housings?
Runs start at 1,000 pieces, with tooling made in-house to your print. We return a quote within 24 hours of the drawing, and first articles are measured on the CMM before the full run so the shell and the weld are confirmed on metal.
Related Products
Request a Quote for This Part
Send a drawing or sketch — get material advice, tooling cost and a per-piece price within 24 hours (working days).




