3-D printing with a MIG welder (“WAAM”) suffers from a positive-feedback problem which impairs geometrical precision: a little bump on one layer tends to attract the arc of the next level and consequently the droplets of metal, becoming a bigger bump on the next level. (Marcin Jakubowicz says, “if you blast the power up, that issue goes away,” but apparently it’s an issue lots of WAAM companies have, and Joshua Pearce reports in the same conference call that he was scrubbing his prints with a wire brush between layers to reduce this problem.)
Adam Blumhagen has proposed to Open Source Ecology that maybe if you run a grinder over the surface after each layer, like the Ability 3D and Big Metal, you could solve this problem (also potentially getting better resolution than what MIG suffers from surface tension). However, an alternative is to stabilize the system with negative feedback: if you detect that the surface is slightly higher, you can compensate by adding less metal to it.
There are a variety of ways you could detect this. MIG welders in particular are not really designed for this; they try to maintain a constant arc length by maintaining a constant voltage across the arc, but the amount of wire stickout is sort of uncontrolled, being the integral of the difference between the meltoff rate (which is nearly proportional to the current, which you can measure) and the wire extrusion rate. A small error in measuring either of these will work out to a large error in estimating the stickout over time.
You could still use the wire as a conductive CMM probe by letting it cool down first, then using a much smaller voltage and current to probe the surface. If you instead periodically probe a known surface that isn’t changing significantly, ideally a piece of graphite or something, you can find out what the current stickout is, correcting the accumulated error in the stickout estimation. Your stickout error will still drift pretty fast, but maybe not fast enough for the positive feedback problem to get out of control.
Using a much thicker electrode, as in stick welding, would largely solve the problem by reducing the linear speed of meltoff; so would using an electrode that isn’t constantly melting off, as in TIG welding, although if you’re constantly crashing your tip into the work it may not retain its nominal geometry for very long, even if you wait for the work and the tip to cool first.