The process argument gets fought on the internet as though it were about quality. Mostly it's about what your rulebook permits, what you can actually reach, and how good you are on the day.
Where the rules decide for you
Start here, because it closes the question for a lot of people.
Sanctioning bodies that permit 4130 chromoly almost universally require TIG for it. MIG is not accepted. So if you're building a chromoly cage, the decision is made and the rest of this is reading for interest.
Mild steel in DOM is normally allowed to be MIG welded. Which is why mild steel is the sensible first-build material even for people who can TIG — not because MIG is better, but because it removes a constraint.
What each process is actually good at
MIG puts down metal fast, tolerates a less-than-perfect fit, and you can run it one-handed. That last part is not a small thing on a chassis. Half the welds on a cage happen with you upside down in a footwell holding the torch at an angle that makes no ergonomic sense, and MIG lets you do that with one hand while the other braces you.
The cost is control. More heat into the joint, more spatter, and a weld that can look acceptable while hiding lack of fusion underneath. The classic MIG failure on tube is a bead that sits on top of the parent metal rather than fusing into it — cold lap — and it will look entirely convincing right up until something loads it.
TIG gives you control over everything independently: heat, filler, travel speed, all separate. Cleaner, less heat input, less distortion, and nowhere for a defect to hide — a bad TIG weld looks bad, which is a safety feature.
The cost is two hands, a foot, good access, genuinely clean material, and a fit-up that's actually tight. TIG punishes a gappy notch far harder than MIG does. Which loops back to the notching article: on thin wall, your fit-up quality and your process choice aren't independent decisions.
Consumables, briefly
For mild steel DOM: ER70S-6 wire on MIG with 75/25 argon-CO₂, or ER70S-2 rod on TIG with straight argon. Nothing exotic.
For 4130, the counterintuitive part is that most chassis builders deliberately use a lower strength filler — ER70S-2 — rather than matching the base metal. The reasoning is that a slightly softer, more ductile weld deposit is less likely to crack than a hard one, and on a thin-wall tubular structure ductility at the joint is worth more than peak strength. ER80S-D2 is the other common answer. People hold strong views; both get used on cars that pass tech.
Tungsten: 2% lanthanated, 1/16" for thin wall, 3/32" when you're on 0.120". Grind it long and pointed for tube work, because you'll be reaching into places where a stubby electrode can't see the joint.
Rough amperage rule: about one amp per thousandth of wall. 0.095" wall wants somewhere near 90–100A. Treat that as a starting point for your first test coupon, not a setting.
The 4130 heat-treatment argument
You'll read that chromoly needs preheat, and that it needs post-weld stress relief, and that it needs neither. Here's the honest state of it.
On the thin wall used in chassis work — 0.083" to 0.120" — most experienced builders don't preheat, and sanctioning bodies generally don't require it. What matters more is not cooling it fast. Don't weld in a draught, don't hit it with compressed air, don't weld a joint and then move straight to quenching the area so you can handle it.
Post-weld normalising of a whole cage is impractical and nobody does it. The parts of this argument that come from aircraft tube practice don't transfer cleanly to a car, because the wall thickness and the service loads are different problems.
If your rulebook says something specific, it outranks all of the above.
The thing that actually determines weld quality
Access.
Not process, not filler, not machine. Whether you can physically get the torch to all the way round the joint with a workable angle and see what you're doing.
This is a design problem, and it's the main reason build sequence shows up in the fabrication output. A joint that's reachable on Tuesday is buried on Thursday once two neighbouring members are tacked. The second-worst weld on any frame is the one you did blind with a mirror. The worst is the one you couldn't reach and quietly left at 70% of its circumference.
So when you're looking at a model and thinking about welding, ask of every joint: where does the torch go, what's behind it, and what has to be welded before this becomes impossible? Then order the build accordingly. That question costs nothing on screen and is unanswerable once there's steel in the way.
Comments
Loading comments…