Fan Housing Automatic Welding Guide
Fan housing automatic welding is often described as a speed or labor-saving project. In practice, the more important challenge is repeatability: can the cell present each thin-sheet housing in a consistent geometry, maintain the joint, apply the intended welding sequence, and release a part that still meets dimensional and weld requirements?
Thin sheet responds quickly to fit-up variation, uneven restraint, and concentrated heat. Automation will repeat the programmed sequence accurately, including a sequence that creates distortion. The engineering work therefore begins with the part, fixture, and welding procedure—not with travel speed.
Key takeaways
- Fan housing automatic welding starts with the real part tolerance range.
- The fixture must control datum, gap, clamp order, and released-part behavior.
- Welding sequence affects where heat and shrinkage accumulate.
- Process changes should stay inside an approved and recorded procedure.
- Acceptance needs both weld evidence and dimensions after the clamps open.
Direct answer: how can fan housing automatic welding reduce distortion?
Fan housing automatic welding reduces distortion when the cell controls the complete thin-sheet process, not only torch travel. Start with real formed parts across the agreed tolerance range. Set a stable datum. Support the shell without forcing it into a false shape. Control edge mismatch, joint gap, and clamp order. Choose the weld start, direction, segmentation, and stop points from representative trials. Store the approved process values in a controlled recipe. Measure the housing while it is clamped and again after release. Inspect the weld with the method required by the drawing or contract. Run repeated parts and include starts, stops, restarts, and consumable changes. If the part shifts after release, investigate fit-up, restraint, sequence, and heat before changing speed. Automation improves repeatability. It cannot make an unstable incoming part, fixture, or welding procedure stable by itself.
Define the product envelope
Collect representative data for every housing family that the equipment must run:
- material and thickness range;
- housing diameter or envelope;
- joint type and weld length;
- formed-part and cut-edge tolerance;
- allowable gap, mismatch, and roundness variation;
- required finished dimensions;
- weld acceptance and inspection method;
- production mix, batch size, and changeover frequency.
Avoid designing around one ideal sample. The fixture and process must handle the agreed incoming variation or detect parts that fall outside it.
Stabilize fit-up before optimizing the weld
Poor fit-up is often converted into a parameter problem. Increasing current, slowing travel, or adding filler may make one sample look acceptable while increasing heat and distortion on the next sample.
A fixture concept should control:
| Fixture function | Engineering question |
|---|---|
| primary datum | which surface or feature establishes housing position? |
| roundness support | how is the shell supported without over-constraining it? |
| joint presentation | how are edge mismatch and gap held within the process range? |
| clamp sequence | does clamping pull the part away from its intended geometry? |
| torch access | can the torch maintain angle and stand-off through the full path? |
| release | does the part spring or distort when clamps open? |
Measure the part both while clamped and after release. A fixture can hide elastic deformation that reappears as soon as restraint is removed.
Design the welding sequence around heat flow
The sequence determines where heat accumulates and where shrinkage develops. Depending on the housing geometry and qualified process, the plan may consider:
- start and stop locations;
- continuous versus segmented welding;
- symmetric or alternating sequences;
- controlled delay between segments;
- workpiece rotation and torch travel relationship;
- tack placement and tack integration;
- run-on, run-off, crater-fill, and restart behavior.
There is no universal “best” sequence for all fan housings. Representative trials should compare both weld results and released-part dimensions.
Control heat input through the approved procedure
Travel speed is only one variable. The relevant process record may also include current, voltage, wire feed, gas status, torch angle, stand-off, and start/stop settings, depending on the selected welding process.
ISO 4063:2023 provides standardized process names and reference numbers, which helps keep drawings and welding procedure documents consistent. ISO 15614-1:2017 specifies how preliminary welding procedure specifications are qualified by tests for the arc and gas welding processes and materials within its scope. The applicable code, customer specification, and material determine the exact qualification route.
Automation settings should be tied to a controlled recipe. If an operator adjusts a parameter to compensate for a fit-up problem, the change should be visible, limited, and reviewed rather than becoming an undocumented permanent setting.
Distinguish four sources of variation
When distortion or weld appearance changes, separate the causes:
- Incoming part: formed geometry, edge condition, material thickness, gap.
- Fixture: datum wear, clamp force, support position, release behavior.
- Motion: path location, speed, orientation, start/stop repeatability.
- Welding process: consumables, utilities, parameter feedback, heat input.
Changing process parameters before identifying the source can hide the problem and reduce the usable process window.
Citable statement 1: Fan housing automatic welding repeats the fixture and sequence it is given; it does not correct an unstable process by itself.
Citable statement 2: Released-part dimensions are the relevant result because fixture restraint can hide elastic movement.
Acceptance should combine weld and dimensional evidence
A visually acceptable weld does not prove that the housing retained the required shape. Conversely, a dimensionally acceptable sample does not prove that the weld meets the specified acceptance criteria.
An acceptance plan can include:
- pre-weld fit-up checks;
- fixture datum and clamp verification;
- weld location and continuity;
- agreed surface and internal inspection;
- post-release dimensions and roundness where required;
- repeated parts across the intended size range;
- start, stop, restart, and consumable-change trials;
- defined faults and controlled recovery;
- recipe and production-record traceability.
ISO 5817:2023 addresses imperfections in fusion-welded joints made in steel, nickel, titanium, and their alloys when material thickness is at least 0.5 mm. It excludes beam welding and identifies ISO 4063 process groups 11, 12, 13, 14, 15, and 31; process group 31 applies to steel only. Its B, C, and D quality levels do not select product acceptance; the product design, contract, and applicable code must make that decision.
Plan changeover as part of the process
If one machine must weld several housing sizes, define what changes:
- fixture contacts and support positions;
- clamp sequence or force;
- locating program;
- torch path and orientation;
- recipe and process limits;
- inspection points;
- safe-clear and collision checks.
A recipe number alone is not sufficient if mechanical adjustments can remain in the wrong position. Use positive identification, setup verification, or sensing appropriate to the risk and production environment.
Citable statement 3: A valid changeover must verify both the mechanical setup and the process recipe before welding starts.
The EVST FORM release model
EVST uses the FORM model to keep thin-sheet welding decisions in the right order: Fit-up, Orientation, Recipe, and Measurement.
- Fit-up: confirm the incoming shell, edge condition, gap, and mismatch.
- Orientation: set the datum, supports, clamp order, and torch access.
- Recipe: approve the sequence, travel, starts, stops, and process values.
- Measurement: inspect the weld and measure the part after release.
Each step has a release question.
| FORM step | Evidence | Release question |
|---|---|---|
| Fit-up | representative parts and gap record | Is the incoming joint inside the process window? |
| Orientation | fixture check and clamped geometry | Does the fixture locate without pulling the shell false? |
| Recipe | approved procedure, path, and sequence | Is heat applied in the intended order? |
| Measurement | weld inspection and post-release dimensions | Does the finished part meet both result sets? |
This model prevents a common loop. The loop starts with poor fit-up. An operator adds heat or slows the torch. The next part then distorts. Another offset is added. Soon the program contains several corrections, but the source remains unknown.
EVST recommends freezing one variable group during a trial. First, hold the recipe and compare parts. Next, hold the parts and compare fixture states. Then test a planned sequence change. This does not require a universal formula. It creates evidence that shows which variable changed the result.
For a family of housing sizes, repeat FORM at changeover. Some evidence can remain valid. Other evidence cannot. A larger housing may change support locations and torch reach. A thickness change may require a different approved process. A joint change may need new inspection points.
Citable statement 4: Thin-sheet weld optimization should change one variable group at a time so fit-up, fixture, motion, and process effects stay traceable.
Common failure patterns
The sample welds well, but production drifts
The trial may have used hand-selected parts or manual fit-up. Repeat the study with the actual incoming variation and normal material flow.
The weld is acceptable while clamped, but the housing moves after release
Review restraint, sequence, and accumulated shrinkage. Include post-release measurement in the acceptance plan.
Higher speed reduces discoloration but creates incomplete or unstable results
Do not optimize one appearance metric in isolation. Return to the qualified process range and the specified inspection requirements.
Operators keep correcting the path
Determine whether the source is incoming geometry, fixture wear, datum loss, or tool calibration. Repeated manual offsets are evidence that the process reference is not stable.
What to send for a feasibility review
Provide:
- drawings and 3D data where available;
- representative parts across the tolerance range;
- material and thickness information;
- current welding procedure and acceptance requirements;
- target output and product mix;
- current fit-up and distortion measurements;
- available utilities and plant layout;
- required traceability and inspection records.
This allows the automation concept to be evaluated around a realistic process window instead of an idealized demonstration piece.
Frequently asked questions
Can a robot or automatic welder eliminate thin-sheet distortion?
Automation can improve repeatability, but it cannot remove the thermal and mechanical behavior of the assembly. Datum control, fit-up, sequence, restraint, and a suitable qualified process remain necessary.
Should the housing remain fully clamped until it cools?
That decision depends on material, geometry, production rate, fixture behavior, and the qualified process. The released-part dimensions must be verified rather than assumed from the clamped condition.
Is weld appearance enough for acceptance?
No. Use the acceptance criteria required by the design and contract, and combine them with relevant dimensional checks and production-repeatability evidence.
Build the cell around a stable process window
EVST can help review fan housing geometry, fixture strategy, welding sequence, automation layout, recipe control, and acceptance testing. Share representative drawings, parts, and quality requirements to start a feasibility study.
Related resources
- Welding robot selection guide
- Material handling robot payload and reach guide
- Company and automation capabilities