What Type of Power Source Is Used for GMAW? A Welder’s Complete Guide (2026)

A young fabricator once bought a decent TIG machine, borrowed a MIG gun from a friend, and tried to run GMAW on the same constant-current output. The arc sputtered, the wire stubbed into the plate, and he burned three nozzles before lunch. Nothing was broken — the machine was simply the wrong type. GMAW lives or dies by its power source, and knowing which one the process actually needs is the difference between a stack of clean beads and a afternoon of frustration.

The Short Answer: GMAW Runs on a Constant-Voltage DC Power Source

If you are looking for a direct answer to the question, “what type of power source is used for GMAW”, it is this: Gas Metal Arc Welding (GMAW), more commonly known as MIG or MAG, runs on a constant-voltage (CV) direct current power source, most often with direct current electrode positive (DCEP). While that sentence sums it up, the reasoning behind it is what makes GMAW one of the most practical and widely used welding processes today.

Why Constant Voltage Matters

Unlike manual processes where you control arc length by moving the electrode, GMAW feeds a continuous wire electrode through the gun at a fixed rate. As you weld, your hand position, joint fit-up and contact-tip-to-work distance (CTWD) are constantly changing. Without correction, these shifts would either burn the wire back into the contact tip or leave it stubbing into the plate. This is exactly where the CV power source comes into play.

A constant-voltage output holds the preset voltage steady, while the welding current rises or falls in response to changes in arc length. If the wire moves closer to the work, resistance drops, current increases and the wire melts faster, pushing the arc back out. If it pulls away, current drops and the wire feeds further into the arc to regain length. This automatic response is known as the self-regulating arc. It is the reason GMAW can be learned quickly and is described by many manufacturers as one of the most forgiving arc welding processes (Lincoln Electric, choosing a welding power source).

No power source is perfectly flat, however. Real GMAW machines have a slight droop in their output characteristic, typically about 2 V drop per 100 A of current increase (NIST IR 3996). This small slope is intentional, as it is flat enough to allow self-regulation, but steep enough to keep the operating point stable during welding.

Power Range and Practical Limits

The amount of power required for GMAW can vary greatly based on wire size and material thickness. Arc welding currents typically range from roughly 100 A to 1000 A, translating to about 2 kVA to 80 kVA of input power depending on the process and duty cycle (NRC welding training module). For GMAW specifically, that range covers everything from light tacking on sheet metal to heavy multi-pass welds on thick plate. Matching your machine’s output to your workload is a key part of proper setup, and you can learn more in How to Set Up a MIG Welder: Complete Step-by-Step Guide.

Polarity and Transfer Considerations

For most applications, DCEP is the standard for GMAW. This polarity concentrates heat at the workpiece, produces good penetration and allows for smooth wire melting with the shielding gases used for carbon steel, stainless steel and most alloys. While DCEN (straight polarity) is far less common, it can be used in specific cases, such as with helium-rich gas on aluminum or certain brazing applications, to achieve a wider, shallower bead. If you want to see how polarity compares across different processes, check out our guide to electrodes, numbers and polarity.

The CV output is what makes the different metal transfer modes possible, from short-circuiting to spray and pulsed spray. These modes are fine-tuned by adjusting voltage, wire feed speed and shielding gas, all while relying on the self-regulating nature of the power source.

Multi-Process Machines and Common Mistakes

Many modern welders are multi-process units. A single inverter can provide CV for GMAW and FCAW, constant current (CC) for stick and lift-arc TIG, and sometimes AC output as well. These machines are incredibly versatile, but the setting you select is critical. Running GMAW on a CC output will remove the self-regulating arc, leading to excessive wire stubbing, spatter and frustration. You can learn more about selecting the correct mode in how MIG/TIG combo welders are actually configured.

In short, the power behind GMAW is not brute force alone. It is the combination of a constant-voltage DC source, a slight droop, and DCEP polarity that allows the machine to correct for your movements in real time. That is what keeps the arc stable, efficient and easy to control, and it is why this answer remains the foundation for every other aspect of the process.

GMAW Fundamentals: Why the Power Source Controls Everything

Before we dig into volt-ampere curves and polarity charts, it helps to be clear about what GMAW actually is. The American Welding Society defines gas metal arc welding as a process that maintains an arc between a continuously fed filler metal electrode and the base metal, with the weld pool shielded by gas supplied from an external source (AWS Welding Digest). Two things in that sentence drive everything else here: the electrode is a moving wire, and something outside the arc has to keep both the wire and the puddle stable while a human hand moves them.

That “something” is the power source. It is not just a box that makes sparks. It decides how fast the wire melts, how deep the arc bites, how much spatter you clean up afterward, and whether a beginner lays a clean bead on their first afternoon or quits in frustration. If you have ever followed a step-by-step MIG welder setup guide, you already know how much of the process comes down to picking the right output characteristics before you strike an arc.

The Welding Circuit: A Closed Loop

GMAW only works when the entire circuit is intact. Current leaves the power source through the welder cable, runs into the gun, passes through the contact tip (also called the contact tube), feeds into the wire electrode, crosses the arc, enters the workpiece, and returns through the work clamp and its cable. The U.S. National Institute of Standards and Technology states it plainly: in GMAW the current passes from the contact tube to the electrode, through the arc, and out to the weld joint (NIST IR 3996).

Every one of those connections is a potential weak link. A loose Dinse plug, a corroded work clamp, a worn gas nozzle, or a galled contact tip adds unwanted resistance and turns a smooth arc into an erratic one. That is why experienced welders treat their cable connectors and grounds as maintenance items, not afterthoughts.

Wire Feed Speed Is Really an Amperage Control

Wire diameters in GMAW commonly run from about 0.6 mm to 1.6 mm (roughly 0.024 in to 0.062 in), loaded on a spool and pushed through the gun by drive rolls. Here is the part that trips up almost every beginner: on a proper GMAW machine, the wire feed speed knob is effectively your amperage knob.

It works because burn-off rate and feed speed chase each other. Push the wire faster and more cold metal enters the arc, momentarily shortening it and driving current up; the arc then melts the wire faster until melt-off exactly matches the feed rate. Slow the feed and the opposite happens. The system settles into equilibrium, so the feed speed you dial in determines the welding current you actually get. Voltage, by contrast, mostly governs arc length and bead width. That division of labor is the foundation of GMAW parameter setting, and it is why a CV machine lets one dial do so much work.

The Real Problem: Arc Length Never Sits Still

In semi-automatic GMAW, the welder holds the gun, so the distance between the contact tip and the workpiece is always moving. That distance is contact-tube-to-work distance, or CTWD, and it equals electrode stickout plus arc length. Even a steady hand drifts a millimetre or two; a joint that changes contour, a corner joint, or an out-of-position overhead pass makes it worse. Every one of those wobbles instantly changes the electrical resistance of the circuit.

This is the physical problem the power source must solve, in milliseconds and without any input from the operator. If the arc gets longer, resistance climbs and current drops; less current means slower melt-off, which would make the arc even longer still. Left alone, the process runs away.

That self-correcting behavior is what makes GMAW forgiving. When the welder unintentionally increases stickout, the added resistance would normally starve the arc, but a CV power source responds by pushing current up to restore balance in fractions of a second. The arc essentially regulates its own length while the machine supplies whatever current the moment demands. It is also why running GMAW on the wrong output, such as a drooping constant-current TIG machine, feels hopeless: nothing self-corrects, and every hand movement lands directly in the puddle.

None of this replaces good fundamentals. Safe operation still means sorting your welding PPE before the arc is struck, and process fundamentals still matter when you move to a wire that makes its own slag, as covered in our flux-core wire welding tips. But once you internalize the loop — feed speed sets current, voltage sets arc length, and the machine compensates for your hand — the next sections on CV and CC characteristics will feel obvious rather than abstract.

Constant Voltage (CV) Power Sources: The Standard for GMAW

If you remember one technical fact from this entire guide, make it this one: GMAW is built around a constant-voltage (CV) DC power source. Not constant current, not a transformer-only AC buzz box. CV. From a $300 hobby unit to a $25,000 robotic cell, nearly every MIG machine ever built holds that flat voltage characteristic at its core. Understanding why CV is the right output is the difference between fighting your machine and letting it work for you.

What \”Constant Voltage\” Actually Means

A constant-voltage source delivers a relatively fixed welding voltage no matter how much current the arc demands. \”Constant\” is a slight exaggeration: the voltage does sag a little as current rises. NIST reference data in Welding Handbook (NIST IR 3996) puts the typical slope at roughly 2 volts per 100 amps. Set your machine to 22 V and draw 200 A, and the output may settle nearer 18–19 V. That small sag is not a flaw. It is the whole mechanism that makes GMAW self-regulating.

Compare that to a constant-current (CC) machine, which holds amperage steady and lets voltage float. CC is the correct answer for TIG on sheet metal, for stick welding, for plasma cutting — any process where the welder’s hand controls the arc length and the machine simply feeds the heat. NRC and AWS training material states the division plainly: constant-voltage sources are generally used with processes that employ a continuously fed consumable electrode — GMAW and FCAW — while constant-current sources serve the manually fed processes.

The Self-Adjusting Arc: How CV Regulates Itself

Here is the physical problem GMAW creates. You feed wire from a spool through a gun at a set speed. The wire melts in the arc, and arc length is simply the gap between the end of the wire and the weld puddle. In semi-automatic welding your hand moves constantly — accelerating, slowing, changing angle, climbing around a joint. Arc length shifts every millisecond. Something has to correct it instantly, and it cannot be you, because you are too slow.

Under CV, the correction happens electrically and automatically. ScienceDirect’s engineering reference on power sources describes the arrangement neatly: in CV GMAW, voltage and wire feed speed are preset, while welding current is controlled by the wire feed speed. Push the feed faster and the arc shortens, the machine pumps out more current, the wire melts faster, and the arc length returns to where you set it. Slow the feed and the opposite happens. The arc is, for all practical purposes, self-adjusting.

That is exactly why beginners pick up MIG in an afternoon and struggle for weeks with stick. On a CV machine you set two numbers and pull the trigger. There is no manual arc-length dance to learn first.

Stickout, Resistance and Instant Self-Correction

The self-regulation shows up most clearly when something changes in the circuit. Stickout — the distance from the contact tip to the work, or CTWD — is the classic example. As electrode stickout increases, the extra length of wire adds electrical resistance, and current falls instantly. Lower current means less heat, so arc voltage and heat input drop with it.

Now watch what the CV source does. With voltage commanded to hold, the machine responds by drawing more current to restore the balance, and the arc recovers on its own. The welder never feels it happen. On a CC machine, that same stickout change would produce a wild voltage swing and an unstable arc the operator has to chase by hand. This behavior is why professional MIG setups obsess over consistent stickout and gun angle — as covered in our complete MIG welder setup guide.

Slope, Open-Circuit Voltage and Why CV Feels Easy

Two practical numbers matter when you read a machine’s spec plate. First, slope: a stiffer CV source holds voltage more tightly, giving crisper short-circuit control. A softer machine sags more, producing a gentler, broader puddle. Most shop machines sit in that 2 V per 100 A neighborhood; premium inverters are flatter still.

Second, open-circuit voltage (OCV) — the voltage present before the arc strikes. CV GMAW machines typically show 50–80 V OCV, enough to reliably initiate the arc through mill scale or light contamination without stalling. Too little OCV and the wire shorts to the plate and stubs; too much raises electrical risk and fume, which is why OSHA’s ventilation and protection rules exist in the first place.

Why the Same CV Machine Also Drives FCAW

There is a useful bonus here. Flux-cored arc welding uses the same continuously fed consumable wire, so the same CV output serves it — which is why flux-core wire welding tips read almost identically to MIG settings guides. Swap the wire, flip polarity to DCEP or DCEN depending on the tubular wire, adjust gas or go gasless, and the power source needs no change. That shared CV backbone is also what makes MIG/TIG combo machines practical: one box, two entirely different arc characteristics, switched electronically.

The bottom line is simple. Constant voltage is not an industry habit or a marketing spec — it is the electrical condition that lets a consumable wire electrode feed, melt and regulate itself thousands of times a minute. Get the output characteristic wrong, and no amount of technique will save the weld.

Constant Current (CC) and Hybrid Outputs: The Exceptions That Prove the Rule

If constant-voltage is the default answer for GMAW, constant current is why that answer deserves a second look. Plenty of welders own multi-process machines that switch between CV and CC, so it pays to know when each output actually helps. Short version: CC is the natural home of stick and TIG welding, but it turns up in GMAW at the edges — thick plate, research-grade pulsed systems, and modern hybrid processes that borrow traits from both.

What a Drooping (Constant-Current) Curve Actually Does

A constant-current power source holds output current close to the value you dial in even as arc voltage swings around. On a volt-ampere graph the curve “droops” steeply: a big change in arc length produces only a small change in amperage. That behavior is exactly what NRC welding training material describes when it notes that constant-current sources are normally used for GTAW/PAW and SMAW, where arc length variations have minimal effect on current.

That property is a gift for a TIG welder hovering tungsten a few millimeters off the plate, or a stick welder whose electrode shortens with every second of the arc. The puddle gets consistent heat no matter how steady your hand is, which is why rod holders and TIG torches feel so forgiving compared with a MIG gun.

Why Plain GMAW Struggles on CC

Now flip the situation to GMAW. The wire feeds at a fixed speed and burns off at a rate set by the current. On a CC machine, if the wire stubs into the puddle and the arc length collapses, current stays roughly where it was — the machine does nothing to correct it. The wire feeds in faster than it melts, piles up, and stalls. If the arc stretches instead, current barely drops, so nothing pulls the wire back toward the work. Self-regulation disappears, and the welder has to manage stickout manually by balancing wire feed speed against burn-off rate. Some systems close that loop by using arc voltage to vary wire feed automatically, but on a plain CC machine the operator is the control system.

The physics back this up. A 2011 Japan Welding Society numerical study of GMAW found that with CV output, arc power changes substantially with arc length, whereas with CC output arc power hardly changes — which sounds stable, but in a consumable-wire process it means the machine can’t sense and correct the disturbance that matters most.

Where GMAW Genuinely Uses CC

CC is not forbidden in GMAW; it’s situational:

• Pulsed GMAW research rigs: laboratory and R&D systems frequently run pulsed waveforms on constant-current output, where peak and background currents are controlled precisely, pulse by pulse.
• Some CC-based GMAW-P systems: certain pulsed processes and older synergic packages regulate current directly and let wire feed or arc voltage handle the length correction.
• Thick-plate work: as ScienceDirect’s engineering topic summary puts it, the constant-voltage source is normally used for thin-sheet welding, while constant current is preferred for thicker plates, where deep, controlled penetration and heat input matter more than arc-length self-correction.

Deep-rooted joints, heavy sections, and some positional applications fall into this camp — the same reason a procedure might specify CC for a multi-pass root on pipe work like that covered in Pipe Welding Procedures: The Complete Guide to WPS, Positions, Passes & Inspection.

The Thin-Sheet vs. Thick-Plate Trade-Off

Thin sheet punishes excess heat: blow-through, distortion, and burnt edges show up fast, so the CV machine’s easy arc starting, fast self-regulation, and gentle control of short-circuit transfer make it the default. Thick plate punishes insufficient penetration: you want stable, high current held at the fusion line, and CC’s indifferent response to arc-length wander gives you that — provided the operator keeps contact tip-to-work distance disciplined. Increasing stickout on CC raises circuit resistance and drags current down, and the machine won’t bounce it back. You will.

Modern Machines: Switching Between Both

Today’s multi-process units rarely commit to a single curve. A modern inverter can present a drooping CC output when you’re running TIG on aluminum, a flat CV output when you pull the MIG trigger, and a processed hybrid output when you select a pulsed program. Processes like STT and Power Mode are technically neither pure CC nor pure CV: arc-voltage feedback adjusts the waveform in real time so the machine controls arc length while the power source still behaves, externally, like a controlled-current device. Aluminum-oriented programs such as Pulse-on-Pulse work the same way, layering a rhythmic peak current for a cleaner, more attractive bead.

If you’re cross-shopping a dedicated MIG box against a combo unit, our MIG TIG Combo Welders: The Complete Guide to Buying and Using One walks through when the dual-output convenience is worth it and when a single-purpose CV machine simply does the job better. And when you set a CC-mode stick pass alongside your GMAW work, parameter references like 3/32 E7018 Amperage: Settings, Polarity & How to Run It show how differently a drooping curve is dialed in compared with a flat one.

Polarity Matters: DCEN vs DCEP in GMAW

Knowing that GMAW runs on a constant-voltage DC power source only gets you halfway to a weldable setup. You also have to decide which way the current flows. Polarity describes the direction of current between the machine, the arc, and the workpiece, and on a DC machine the electrode sits on either the negative or the positive terminal. That single choice reshapes the heat balance in the arc, and it shows up immediately in penetration, bead shape, spatter, and how cleanly the arc cuts through surface oxides.

The two settings carry a few sets of names, which is where beginners get confused:

• DCEN (direct current electrode negative) is also called straight polarity. The wire is connected to the negative terminal and the workpiece to the positive.
• DCEP (direct current electrode positive) is also called reverse polarity. The wire is connected to the positive terminal and the workpiece to the negative.

Where the heat goes

In a welding arc, most of the heat develops at the anode, the positive member of the circuit. Electrons stream away from the cathode and slam into the anode, converting that kinetic energy into heat. Because polarity decides which member is the anode, it decides whether the arc heat lands mainly on your filler wire or mainly on the base metal.

Why DCEP dominates GMAW

Wire welding is built around DCEP. The large majority of GMAW applications are run electrode positive because it melts the filler wire efficiently and predictably while giving the operator a quiet, stable arc across the short-circuit, globular, and spray transfer modes. The extra heat into the wire end lets you dial deposition with voltage and wire feed speed rather than fighting an erratic arc.

DCEP also puts the workpiece on the negative side, and the cathode is where the cleaning action happens. Electrons erupting off the work surface blast away the thin oxide layer that sits on top of it. On aluminum and stainless, that scrubbing effect is what keeps the weld from wandering or undercutting. For a practical look at how this plays out on real joints, see our guide to aluminum MIG welding, where polarity choice is often the difference between a wetted bead and a pile of ball-shaped spatter.

When DCEN earns its place

Straight polarity is not wrong, it is simply specialized. DCEN pushes arc energy into the base metal, so it can be useful for heavier sections where penetration matters more than bead cosmetics, and it is common in high-deposition globular transfer work run in the flat position. The trade-offs are a less stable arc, a narrower and more rippled bead, and virtually no oxide cleaning, which is why it is a poor choice on aluminum. Outside of GMAW, polarity habits differ by process altogether: flux-cored arc welding often runs electrode negative, which catches people who switch between wires without checking the leads.

Reading the machine and the wire

Most dedicated MIG power sources are hard-wired electrode positive internally, so the polarity is effectively fixed and no field change is needed. Multi-process units and stick machines that share a GMAW gun usually bring both terminals to the front panel, sometimes with a plug-in reversing jumper. When in doubt, consult the manufacturer’s welding basics documentation or the American Welding Society terminology, since electrode positive and reverse polarity mean exactly the same thing and mixing them up is the fastest route to a confusing arc.

Testing is simple: run a short bead on scrap of the same thickness and look at it. An arc that crackles, spatters heavily, and leaves a tall narrow profile usually means you are on the wrong polarity for the wire. A smooth, flat bead with a quiet buzz and even tie-in at both toes tells you the circuit is right.

Let me count the current fragment’s words roughly and see if it’s too long. The task: rewrite between 640 and 960 words. Let me estimate.

Let me count the current fragment words. I’ll estimate section by section.

Heading: 6 words.

Para 1: ~95 words.
Para 2: ~30 words.
H3: 5 words.
Table: headers + 4 rows ~ let’s count: header row ~12 words; each row ~ 25-30 words. Total table ~110 words.
H3: ~10 words.
Short-circuit para 1: ~85 words.
Haynes para: ~80 words.
Callout: ~40 words.
H3 spray: ~6 words.
Spray para 1: ~65 words.
Spray para 2: ~90 words.
H3 globular: ~5.
Globular para: ~60.
H3 pulsed: ~8.
Pulsed para 1: ~65.
Pulsed para 2: ~70.
H3 dialing in: ~8.
Paragraph: ~15.
List: ~65.
Final para with links: ~90.
Note callout: ~55.
Lincoln para: ~45.

Total ~ 1,150-1,250 words. So it’s too long; need to cut to 640-960. Let’s target ~880 words.

Wait — but does the table count toward word count? Probably yes to be safe. Let me target around 850-900 words including table.

I need to keep the same facts and all links. Let me list links:
1. AWS: https://www.aws.org/magazines-and-media/welding-digest/wd-june-2025-what-is-gmaw
2. Haynes International: https://haynesintl.com/en/alloys/welding-and-fabrication/welding-and-joining/gas-metal-arc-welding-gmaw-mig/ (appears twice in original — but “all of its links” — I should include each link once; the rule says weave internal links once each. For external, keeping the link once is fine. Actually “keep … all of its links” — I’ll keep the Haynes link (at least once), NIST, Journal of Physics D, Lincoln Electric.
3. Internal: https://weldgallery.com/how-to-set-up-a-mig-welder-complete-step-by-step-guide/ (anchor “How to Set Up a MIG Welder”)
4. https://weldgallery.com/mig-tig-combo-welders-the-complete-guide-to-buying-and-using-one/ (MIG TIG Combo Welders)
5. https://weldgallery.com/welding-connectors-explained-how-to-choose-install-and-troubleshoot-cable-connectors-that-never-let-you-down/ (welding connector)
6. https://weldgallery.com/flux-core-wire-welding-tips-pro-techniques-for-cleaner-stronger-welds/ (flux core and gas setups)
7. https://weldgallery.com/what-does-a-welder-wear-the-complete-welding-ppe-guide/ (welding PPE guide)
8. https://nvlpubs.nist.gov/nistpubs/Legacy/IR/nistir3996.pdf (NIST IR 3996)
9. https://iopscience.iop.org/article/10.1088/1361-6463/aaf588 (Journal of Physics D)
10. https://www.lincolnelectric.com/en-SG/Welding-and-Cutting-Resource-Center/Welding-How-Tos/Choosing-a-Welding-Power-Source (Lincoln Electric)

I’ll keep each link once. Haynes maybe twice in original — I’ll use once to save words, that’s fine (“keep all of its links” — one occurrence of the URL counts).

Hmm, but careful: if I remove the second Haynes mention, the spray parameters facts remain but the source attribution is only in the short-circuit paragraph. Safer to keep the fact but cite it once. Actually to keep the fact sourced, I could keep the Haynes link in the spray section and mention it in short-circuit without a link? No — better to link once and say “Haynes International publishes…” once. Let me put the link where both parameters are introduced… I could reference the same source with the link on the short-circuit paragraph and then say “Haynes’ typical spray parameters are…” without a second link. That’s fine.

Now write concisely, cutting filler. Target ~870 words.

Let me draft.

—

How the Power Source Shapes Metal Transfer Modes

Here is where the power source stops being an abstraction and shows up in the puddle. GMAW transfers metal across the arc in four recognizable modes — short-circuiting, globular, spray and pulsed spray — and the machine’s output decides which one you get. As the American Welding Society describes the process, the operator selects the mode mainly through arc voltage, with wire diameter and shielding gas doing the rest of the steering.

That explains why a constant-voltage source fits GMAW so well: you dial a voltage, the machine holds it, and the transfer mode holds with it.

table…

Let me count table words carefully. Header: “Transfer mode | Typical arc voltage | What it looks like | Best suited for” = 12 words. Row 1: “Short-circuiting | 16–22 V | Wire touches, shorts, melts off 20–200 times per second with a soft, quiet arc | Thin sheet, out-of-position, root passes” ≈ 22 words. Row 2: ≈ 20. Row 3: ≈ 24. Row 4: ≈ 25. Table total ≈ 103.

Hmm, that’s a chunk. Keep table but maybe trim cell text slightly while keeping facts. Facts must be kept (voltage ranges). I can trim descriptions a bit.

Let me write the whole thing and count.

Draft:

How the Power Source Shapes Metal Transfer Modes

Here is where the power source stops being an abstraction and starts showing up in the puddle. In GMAW, metal crosses the arc in one of four modes — short-circuiting, globular, spray and pulsed spray — and the machine’s output decides which one you get. As the American Welding Society describes it, the operator selects the transfer mode mainly through arc voltage, with wire diameter and shielding gas doing the rest of the steering.

That single fact explains why a constant-voltage source fits GMAW so well: you dial a voltage, the machine holds it, and the transfer mode holds with it.

The Four Transfer Modes at a Glance

Short-Circuiting Transfer: The CV Machine’s Sweet Spot

Short-circuit transfer is where most hobbyists and light fabricators live, and it is the clearest demonstration of why GMAW wants a CV output. The wire stubs into the pool, current spikes, the stub melts off, the arc re-ignites — thousands of times a minute. Because a constant-voltage source reacts instantly to that short, the arc self-corrects without the operator chasing it. The result is a soft arc you can push into a corner, weld overhead, or use to tack sheet without blowing through.

Haynes International publishes typical parameters that give you a realistic starting point: for 0.035 in wire on short-circuit, run roughly 150–200 ipm wire feed, 70–90 A and 18–20 V, on a blend such as 75% argon / 25% helium. Those numbers line up closely with what our How to Set Up a MIG Welder guide recommends for sheet-metal work.

Spray Transfer: Same Machine, Beefier Output

Spray needs two things a small CV machine sometimes can’t give: enough current to keep the wire molten at the tip, and an argon-rich gas (usually 80–95% argon for carbon steel) so droplets pinch off electromagnetically instead of by surface tension. Haynes’ typical spray parameters are 0.045 in wire at 250–350 ipm, 190–250 A, 28–32 V on 100% argon.

Notice the jump — roughly double the amperage of short-circuit. That is why spray is described as flat-and-horizontal work: the puddle is fluid and arc force is high. It is also why buying decisions hinge on output range; a machine peaking at 180 A cannot run true spray with 0.045 wire. NIST’s welding research documents how droplet detachment in this regime is governed by electromagnetic force rather than contact with the plate.

Globular: The Awkward Middle

Globular sits between the two — voltage high enough for a free-flight arc, current too low for fine spray, so heavy globules cross with noticeable spatter. Deposition on thicker flat steel is good, but most operators skip it in favor of pulse. If you see chunky spatter and irregular beads after bumping voltage, you landed in globular rather than spray.

Pulsed Spray: Spray Fusion at Low Average Current

Pulsed GMAW is where the power source does its most sophisticated work, swinging between a high peak current that detaches exactly one drop and a low background current that keeps the arc alive. Our example of 0.035 in wire at a 75–150 A average with 30–34 V delivers spray-like penetration while keeping heat input low enough for thin material and out-of-position work.

That is genuinely different electrical behavior from a plain CV dropper — the output is waveform-shaped, not flat. For more on where that technology sits, see our MIG TIG Combo Welders guide. The underlying arc physics of initiation and droplet detachment is covered in published research in the Journal of Physics D.

Dialing It In: A Practical Starting Point

Use these Haynes-derived windows as your baseline, then adjust by feel:

• Thin sheet, all positions: 0.030–0.035 wire, 18–20 V, short-circuit, 75/25 Ar/He or 75/25 Ar/CO₂.
• Medium plate, flat/horizontal: 0.035–0.045 wire, 24–28 V, globular-to-spray, 90/10 Ar/CO₂.
• Thick plate, high deposition: 0.045 wire, 28–32 V, 190–250 A, 100% argon for true spray.
• Out-of-position spray-like fusion: 0.035 wire pulsed, 30–34 V, 75–150 A average.

Before turning dials, check the gun liner, contact tip bore and work clamp — a worn welding connector or contaminated tip will sabotage correct parameters. The same goes for gas flow: check your flux core and gas setups for leaks before blaming the power source. And because pulsed and spray arcs throw more UV, confirm your PPE with our welding PPE guide.

Lincoln Electric’s power-source selection guidance puts it plainly: match the process, material range and required output first, then let the transfer mode follow. Get that ordering right and the arc tells you the rest.

—

Now count. Let me count carefully-ish.

Heading: 7 words.

P1: “Here is where the power source stops being an abstraction and starts showing up in the puddle.” (17) “In GMAW, metal crosses the arc in one of four modes — short-circuiting, globular, spray and pulsed spray — and the machine’s output decides which one you get.” (~27) “As the American Welding Society describes it, the operator selects the transfer mode mainly through arc voltage, with wire diameter and shielding gas doing the rest of the steering.” (~28) Total ~72.

P2: “That single fact explains why a constant-voltage source fits GMAW so well: you dial a voltage, the machine holds it, and the transfer mode holds with it.” ~28.

H3: 6.

Table: header 12. R1: Short-circuiting(1) 16–22V(1) “Wire touches, shorts and melts off 20–200 times per second; soft, quiet arc” (~14) “Thin sheet, out-of-position, root passes”(4) = 20. R2: 1+1+ “Large globules, heavier spatter, gravity-assisted”(5) + “Thicker carbon steel where spatter is tolerable”(7) = 14. R3: 1+1+ “Fine droplets stream continuously; no contact with the workpiece”(9) + “Flat and downhill work on medium-to-thick steel”(8) = 19. R4: “Pulsed spray (GMAW-P)”(3)+”30–34 V typical”(3)+ “Background current holds the arc; a peak pulse detaches one drop per pulse”(13)+ “Out-of-position spray, aluminum, thin-to-medium material”(6) = 25. Table total ≈ 90.

H3 short-circuit: 8.
P1: “Short-circuit transfer is where most hobbyists and light fabricators live, and it is the clearest demonstration of why GMAW wants a CV output.”(25) “The wire stubs into the pool, current spikes, the stub melts off, the arc re-ignites — thousands of times a minute.”(21) “Because a constant-voltage source reacts instantly to that short, the arc self-corrects without the operator chasing it.”(18) “The result is a soft arc you can push into a corner, weld overhead, or use to tack sheet without blowing through.”(22) ≈ 86.

P2: “Haynes International publishes typical parameters that give you a realistic starting point: for 0.035 in wire on short-circuit, run roughly 150–200 ipm wire feed, 70–90 A and 18–20 V, on a blend such as 75% argon / 25% helium.”(~36) “Those numbers line up closely with what our How to Set Up a MIG Welder guide recommends for sheet-metal work.”(19) ≈ 55.

Callout: “Quick Note: Short-circuit is the easiest mode to dial in because the CV machine corrects for you — raise voltage slightly for lack of fusion, lower it slightly for spatter.” ≈ 31.

H3 spray: 7.
P1: ~ “Spray needs two things a small CV machine sometimes can’t give: enough current to keep the wire molten at the tip, and an argon-rich gas (usually 80–95% argon for carbon steel) so droplets pinch off electromagnetically instead of by surface tension.”(45) “Haynes’ typical spray parameters are 0.045 in wire at 250–350 ipm, 190–250 A, 28–32 V on 100% argon.”(21) ≈ 66.

P2: “Notice the jump — roughly double the amperage of short-circuit.”(10) “That is why spray is described as flat-and-horizontal work: the puddle is fluid and arc force is high.”(19) “It is also why buying decisions hinge on output range; a machine peaking at 180 A cannot run true spray with 0.045 wire.”(25) “NIST’s welding research documents how droplet detachment in this regime is governed by electromagnetic force rather than contact with the plate.”(22) ≈ 76.

H3 globular: 5.
P: “Globular sits between the two — voltage high enough for a free-flight arc, current too low for fine spray, so heavy globules cross with noticeable spatter.”(26) “Deposition on thicker flat steel is good, but most operators skip it in favor of pulse.”(16) “If you see chunky spatter and irregular beads after bumping voltage, you landed in globular rather than spray.”(19) ≈ 61.

H3 pulsed: 7.
P1: “Pulsed GMAW is where the power source does its most sophisticated work, swinging between a high peak current that detaches exactly one drop and a low background current that keeps the arc alive.”(33) “Our example of 0.035 in wire at a 75–150 A average with 30–34 V delivers spray-like penetration while keeping heat input low enough for thin material and out-of-position work.”(33) ≈ 66.

P2: “That is genuinely different electrical behavior from a plain CV dropper — the output is waveform-shaped, not flat.”(18) “For more on where that technology sits, see our MIG TIG Combo Welders guide.”(14) “The underlying arc physics of initiation and droplet detachment is covered in published research in the Journal of Physics D.”(20) ≈ 52.

H3 dialing: 7.
P: “Use these Haynes-derived windows as your baseline, then adjust by feel:”(11)

List: item1 ~13, item2 ~12, item3 ~14, item4 ~11 = 50.

P: “Before turning dials, check the gun liner, contact tip bore and work clamp — a worn welding connector or contaminated tip will sabotage correct parameters.”(25) “The same goes for gas flow: check your flux core and gas setups for leaks before blaming the power source.”(20) “And because pulsed and spray arcs throw more UV, confirm your PPE with our welding PPE guide.”(18) ≈ 63.

Callout: “Note: Transfer mode is not a setting you press — it emerges from voltage, wire feed speed, wire diameter and gas. Change one and you may silently shift modes, which is why one-variable-at-a-time tuning still works best.” ≈ 38.

Final P: “Lincoln Electric’s power-source selection guidance puts it plainly: match the process, material range and required output first, then let the transfer mode follow.”(25) “Get that ordering right and the arc tells you the rest.”(11) ≈ 36.

Total: 7+72+28+6+90+8+86+55+31+7+66+76+5+61+7+66+52+7+11+50+63+38+36.

Sum: 7+72=79; +28=107; +6=113; +90=203; +8=211; +86=297; +55=352; +31=383; +7=390; +66=456; +76=532; +5=537; +61=598; +7=605; +66=671; +52=723; +7=730; +11=741; +50=791; +63=854; +38=892; +36=928.

Hmm 928 — within 640-960 but close to max. Word counting varies; table markup might be counted as words too. Let me trim to ~880 to be safe. Cut some filler.

Actually if counting includes HTML tags as “words” (unlikely for a word counter, but a strict counter counting whitespace-separated tokens would include tags like