Welding Knowledge Questions and Answers for Design Engineers
A working set of welding knowledge questions answers, written for engineers who specify joints, review drawings and hand parts to a machine shop. Each answer covers what the number means on the floor and when a process is the wrong choice.

What This Page Covers
Ten shop-floor topics, from grain structure to inspection, plus a short FAQ at the end.
What Happens Inside the Weld Pool
Weld pool solidification follows the same rules as any liquid metal. Nuclei form and grains grow. What separates a weld from a casting is the substrate: the partially melted grains of the parent metal at the fusion boundary act as ready-made seeds, so grains grow epitaxially from the base metal into the pool instead of starting fresh. That is why a weld's grain structure is continuous with the parent metal and why the heat-affected zone matters as much as the weld itself.
The shape of the pool sets the grain direction. A wide, shallow pool cools slowly and produces coarse columnar grains that run sideways. A narrow, deep pool cools fast and leaves finer grains, which usually means better toughness. Travel speed, current and joint geometry all change that shape, so two welds made with the same filler on the same alloy can behave differently in service.
Solute redistribution is the part most drawings ignore. As the pool freezes, alloying elements and impurities are pushed ahead of the solidification front, which is how segregation bands and centerline porosity form. A joint that passes visual inspection can still have a weak centerline. If the part sees fatigue or pressure, that is the region to check with radiography or sectioning.
Heat Input, Cooling Rate and the Heat-Affected Zone
Heat input controls most of what goes wrong after welding. It is roughly voltage times current divided by travel speed, expressed in kJ/mm. Raise the current or slow the travel and heat input climbs. That softens some alloys, hardens others, and widens the heat-affected zone in both cases.
The heat-affected zone is the parent metal that did not melt but was hot enough to change. In carbon steel, fast cooling can produce martensite and hardness above 350 HV, which is where cold cracking starts. Preheat and controlled cooling are the usual countermeasures. In 6061-T6 aluminium, the same zone loses temper and drops from roughly 310 MPa yield to about 115 MPa near the weld.
A practical rule for the shop: match heat input to section thickness, not to the welder's habit. Thin sheet needs low current and fast travel or it warps. Thick sections need preheat to slow cooling and avoid hydrogen cracking, and interpass temperature has to stay inside the qualified range for the procedure.
Why Parts Move and How to Limit It
Distortion is thermal expansion meeting restraint. The weld shrinks as it cools, the surrounding cold metal holds it back, and the part bends or bows. Three things drive the amount: total weld volume, joint restraint and how the passes are sequenced.
Reduce weld volume first. A 6 mm fillet where a 4 mm one passes the calculation removes a large share of the shrinkage. Then balance the sequence. Backstep welding, balanced passes on both sides of a neutral axis, and skip welding all spread heat instead of concentrating it in one run.
Fixturing helps, but it is not free. Heavy restraints hold the part flat during welding and then release residual stress when the clamps come off, so the part can move during machining. Stress relief before finish machining is the usual fix. For aluminium, that means a controlled thermal cycle rather than a simple bake.
Which Welding Process Fits Which Job
Selection depends on thickness, alloy and how much post-weld machining follows.
| Process | Typical thickness | Best fit | Main limit |
|---|---|---|---|
| TIG (GTAW) | 0.5–6 mm | Thin stainless, aluminium, root passes | Slow; needs clean joint faces |
| MIG (GMAW) | 3–20 mm | Steel frames, production runs | More spatter; less control on thin sheet |
| FCAW | 6–25 mm | Heavy steel, outdoor work | Slag removal; hydrogen pickup risk |
| Laser | 0.5–4 mm | Fine joints, low distortion | Tight fit-up; high equipment cost |
| Resistance spot | 0.5–3 mm | Sheet metal assemblies | Lap joints only; electrode wear |
Alloy Weldability and the Carbon Equivalent
Weldability is not a single number. It is how the alloy responds to the thermal cycle: whether it cracks, how much it loses in strength, and whether the joint stays sound under restraint. Carbon equivalent is the quick screen for steel. A common form is CE = C + Mn/6 + (Cr+Mo+V)/5 + (Ni+Cu)/15. Above roughly 0.40, preheat becomes mandatory, and the required temperature rises with thickness.
Stainless steels behave differently. Austenitic grades such as 304 and 316 weld well but are prone to sensitization if they sit in the 450–850 °C range too long, which pulls chromium out of solution and leaves the grain boundaries open to intergranular corrosion. Low-carbon grades like 316L reduce that risk. Ferritic and martensitic grades need preheat and often post-weld heat treatment.
Aluminium brings its own problems. The oxide layer melts near 2,050 °C while the metal below melts near 660 °C, so the oxide has to be removed mechanically or by AC cleaning action before the pool forms. Thermal conductivity is about five times that of steel, which pulls heat away fast and demands higher current. Hot cracking is a real risk in 6061 filler selection; 4043 and 5356 behave differently and the choice depends on service temperature and strength target.
Inspection Methods and What They Actually Find
Visual inspection catches more than people expect: undercut, overlap, crater cracks, incomplete fusion on the face. It costs almost nothing and should happen before any other method. What it cannot see is below the surface.
Dye penetrant finds surface-breaking defects on non-porous materials. Magnetic particle finds surface and near-surface flaws in ferromagnetic steel only. Ultrasonic testing finds planar defects such as lack of fusion and cracks deep in thick sections, but it needs a trained operator and a reference standard. Radiography shows porosity and inclusions well, but planar defects aligned with the beam can be missed.
Pick the method from the failure mode you care about, not from habit. Pressure vessels need volumetric testing. Fatigue-loaded brackets need crack detection at the toe. A welded frame that only carries static load may need nothing beyond visual and a dimensional check. Over-specifying inspection adds cost without adding safety.
Machining Welded Assemblies
Welded parts come to a machine shop for two reasons: to hit a tolerance the weld cannot hold, and to clean up the joint for a mating surface. Both need planning. The weld is harder than the parent metal in most steels, so tool choice changes. Interrupted cuts across the weld toe are hard on carbide unless the feed and speed are dialed back.
Residual stress is the bigger issue. A welded bracket that is machined flat and then released can bow by tenths of a millimeter. Rough machine, stress relieve, then finish machine is the sequence that holds tolerance. For parts that cannot be heat treated, symmetric material removal on both sides of the weld keeps the stress balance closer to neutral.
Fit-up before welding decides how much machining is needed after. A joint with a 1 mm mismatch costs far more to clean up than one held to 0.2 mm. If the drawing calls for a machined face at the weld, say so on the drawing so the welder knows to leave stock.
Common Questions on Welding and Post-Weld Machining
Can you machine a part after it has been welded?
Yes. Welded assemblies are machined routinely, but the sequence matters. Rough machining before stress relief and finish machining after is the reliable route for tight tolerances.
For aluminium, the heat-affected zone is softer than the base metal, so the cutter will behave differently on either side of the weld line. Expect to adjust speeds and feeds.
What tolerance can a welded assembly hold without machining?
A welded steel frame typically holds ±1 mm or looser, depending on size and restraint. Distortion from shrinkage and fit-up variation stacks up fast.
If the drawing needs ±0.05 mm, plan for a machining operation after welding. No welding process holds that on its own.
Which stainless grade is easiest to weld?
304L and 316L are the most forgiving of the common grades. The low carbon content reduces sensitization risk and the austenitic structure tolerates thermal cycling well.
Martensitic grades such as 420 and 440C need preheat and post-weld heat treatment, and they crack more readily under restraint.
Does preheating really change the result?
It changes the cooling rate, and cooling rate controls hardness in the heat-affected zone. Slow cooling gives softer, more ductile structures and lowers the risk of hydrogen cracking.
For a 25 mm carbon steel section with a carbon equivalent above 0.40, skipping preheat is a gamble, not a shortcut.
How do you weld aluminium without porosity?
Clean the joint mechanically, keep the filler dry, and use the right shielding gas flow. Aluminium oxide holds moisture and that moisture becomes porosity in the pool.
AC TIG with a balanced cleaning action works well on thin sections. For thicker parts, MIG with a pulsed current gives better control of the pool.
What should be on the drawing for a welded part?
Weld symbol, process, filler class, size, and whether the weld is continuous or intermittent. Add the inspection method and acceptance level if the joint is structural.
Note any surface that will be machined after welding and the stock to leave. That one line saves a rework cycle.
Send Us the Welded Assembly
Upload the drawing and we will review weld prep, machining sequence and tolerance stack before quoting.
12-hour quote100% inspectionNDA on request