Skip to content
Home Articles Standards About Contact
MIG Field Study

MIG Welding Vertical Up: The Complete Technique Guide

NH
Nicholas Harper · Jun 16, 2026 · 69 min read
Table of Contents

The first time you run a bead up a vertical plate, gravity wins in about two seconds — the puddle sags, the toes undercut, and you are left with a lumpy caterpillar instead of a weld. Every welder who has chased that molten puddle uphill knows the frustration, and most never get a straight answer about why it happens or how to fix it. This guide explains the physics, the settings, and the hand motions that turn vertical-up MIG from a fight into a controlled process.

Why Vertical-Up MIG Welding Separates Beginners From Professionals

Picture this. You have built a trailer ramp in the driveway, clamped it upright because the hinges are easier to reach that way, and you strike an arc about four inches off the ground. For two seconds everything looks fine. Then the puddle starts to sag. It swells, the toes undercut, and a soft ripple of molten metal sags downhill faster than you can chase it. What rolls out when you stop is a lumpy grey caterpillar that looks nothing like the beads you run on the bench. You turn the machine up. It gets worse. You turn it down, move faster, and the wire starts stubbing and popping.

Almost every welder has been there, and the frustrating part is how rarely anyone explains what actually went wrong. Vertical-up MIG is not simply flat welding with the gun tilted. It is a different negotiation with gravity, and the rules that work on a horizontal plate will actively sabotage you on a vertical one.

What Vertical-Up Actually Means

Vertical-up means exactly what it sounds like: the weld progresses from the bottom of the joint toward the top, with the weld axis running roughly vertical and the weld face lying in an approximately vertical plane. In welding position terminology, that is uphill progression. The distinction matters because running the same joint downhill, from top to bottom, produces a completely different weld with different strengths and different defects.

The core idea is simple to state and harder to master. Because gravity pulls the molten pool downward, you use a shorter arc, less heat, and a small, tightly controlled puddle, then let each pass solidify quickly enough to support the next one. Done correctly, the still-hot metal below acts as a shelf, and gravity actually helps drive the arc into the joint for deep penetration. Done incorrectly, gravity simply drains the pool out of the joint before it can fuse.

As a working rule of thumb, most guides and equipment manufacturers, including Miller’s MIG fundamentals resources, recommend dialing voltage and amperage down by roughly 10 to 15 percent compared with your flat-position settings when you move to a vertical joint. That single adjustment is where most beginners skip ahead and pay for it later.

Who Needs This Skill

This guide is written for the people who actually find themselves welding out of position: home welders fixing equipment in place instead of on the bench, fabrication shop hands running structural frames, apprentices working toward their first performance qualification, and maintenance techs who cannot unbolt the machine to lay it flat. If your joint can be repositioned, reposition it. Flat welding is faster, easier, and cheaper. Vertical-up is what you do when it cannot be moved, or when the joint needs the penetration that uphill progression provides.

That brings us to the honest trade-off. Vertical-up is slower and far less forgiving than flat welding. Travel speeds drop, the margin for error shrinks, and a steady hand matters more than a powerful machine. In return, you get noticeably deeper penetration on thicker steel, better control of the root pass, and a weld that codes and inspectors generally prefer on structural work. Vertical-up technique is typically the recommended approach for materials around 1/4 inch (6 mm) and heavier, where a downhill bead would sit on top of the joint rather than in it.

Quick Note: These settings and techniques in this guide are aimed at solid wire GMAW on mild steel. Aluminum behaves very differently because of its thermal conductivity and oxide layer. If you are welding aluminum, start with the transfer mode and gas choices in our Aluminum MIG Welding: The Complete Practical Guide before adapting anything here.

What You Will Learn Here

We will start with the position system, so you can read a welding procedure specification or a test record without guessing what 3G and 3F mean. In AWS D1.1 structural codes, a 3G qualification covers a vertical groove weld on plate and 3F covers a vertical fillet weld, and the direction of progression, up or down, is recorded as an essential variable on the welding procedure specification. A welder qualified flat is not qualified vertical, and a welder qualified vertical-up is not automatically qualified vertical-down.

From there we move through the physics of what the puddle is doing, machine setup and transfer mode selection, the step-by-step hand technique, a joint-by-joint playbook for grooves, fillets, laps and pipe, a defect troubleshooting table, an honest comparison against vertical-down and flat, the safety and qualification items most guides skip, and a four-week practice plan you can run on scrap.

By the end, vertical-up will still be slower than flat welding, and it should be. But it will stop being a fight.

Quick Reference: The Core Rules

  • Progress from the bottom of the joint upward, never downhill, unless the procedure says otherwise.
  • Reduce voltage and wire speed about 10 to 15 percent from your flat settings as a starting point.
  • Keep the puddle small and the weave narrow, no wider than about 2.5 to 3 times the wire diameter.
  • Pause at each toe long enough to fuse, move quickly across the center.
  • Watch the puddle, not the arc. The puddle tells you what to do next.

Vertical-Up in Plain English: Positions, Terms, and Where It Fits

Position codes look like secret shop shorthand until someone hands you a WPS and asks which test you are qualified for. The system is simpler than it appears: the number tells you where the weld sits in space, and the letter tells you what kind of weld it is. Once you can decode that combination, a test record stops being paperwork and starts telling you exactly what a welder has proven they can do.

Reading the Position Code Without Guessing

In the AWS system, the number runs from 1 to 4: flat, horizontal, vertical, and overhead. The letter identifies the weld type. G means groove weld, while F means fillet weld. For fillets, the letters also describe the position directly: F for flat, H for horizontal, V for vertical, and OH for overhead. That is why you see both 3G and 3F on structural qualification papers.

The code describes the weld axis and the weld face orientation, not your body position or gun angle. The weld axis is the imaginary line running lengthwise through the weld. The weld face is the exposed surface. In a vertical groove weld, the axis runs up and down and the face lies in a vertical plane. In a flat groove weld, the axis is horizontal and the face points upward.

Position Groove / Fillet What it means Typical application
1 1G / 1F Flat; weld face points up Bench fabrication, shop work
2 2G / 2F Horizontal; weld axis runs sideways Horizontal seams and T-joints
3 3G / 3F Vertical; weld axis runs up and down Column splices, stiffeners, vertical tests
4 4G / 4F Overhead; weld face points down Understructure and repair work

Quick Note: The number tells you where the weld sits. The letter tells you what kind of weld it is. Neither one tells you which direction the gun should travel.

What 3G and 3F Actually Test

A 3G test is a vertical groove weld, and a 3F test is a vertical fillet weld. Both are common qualification tests for structural work because they combine a demanding position with a joint that reveals fusion problems quickly. The Minnesota DOT field welder qualification records, for example, list 3G as vertical groove and 3F as vertical fillet, with vertical progression recorded as up.

Progression is separate from position. Uphill and downhill are not extra position numbers; they describe how the arc travels along the vertical axis. That distinction matters because AWS D1.1 treats position and progression as essential variables, and its WPS form includes vertical progression as its own line item. A procedure written for 3G uphill is not automatically interchangeable with 3G downhill.

Why a Flat-Qualified Welder Is Not Vertical-Qualified

Codes treat position as an essential variable because gravity changes the welding process. A welder who can run a sound flat bead has demonstrated wire speed, voltage, and travel control, but not puddle support. Vertical-up welding demands a smaller molten pool, different heat input, and deliberate manipulation to keep the bead from sagging. That is a separate skill, so a flat qualification does not carry over to vertical.

The same logic applies to pipe. 5G is a horizontal fixed pipe position, and 6G is a pipe fixed at roughly 45 degrees. Both include segments that must be welded vertical-up, which is why pipe tests expose weak puddle control so effectively. If you also run stick, our all-position welding rod guide explains how the same position logic carries across processes.

Uphill, Downhill, Groove, and Fillet: Clearing Up the Confusion

Vertical-up and uphill mean the same thing: the gun or electrode travels from the bottom of the joint toward the top. Vertical-down and downhill mean the opposite. These are progression terms. A vertical groove weld is a butt-type joint tested as 3G, while a vertical fillet is a T-joint or lap-joint weld tested as 3F. Both are vertical, but they are not the same test, and they do not behave the same way under the arc.

The practical takeaway is to read a qualification record literally. “3G up” tells you the position, joint type, and progression. It does not tell you the plate thickness, backing, or consumable, which is why the rest of the WPS still matters.

Note: Before you practice for a test, confirm whether the procedure calls for uphill or downhill progression. Practicing the wrong direction can build habits that fail the qualification even when your bead looks good.

We need to rewrite fragment to fit 560-840 words, keeping same heading, facts, links. Need count words. Current fragment likely ~750? Let’s count approximate. It seems around 760 words maybe. But truncated table. We need output full corrected fragment incl table? The current fragment was truncated in prompt at table. Need maintain table likely. Need reconstruct table logically. Hmm. We need produce HTML fragment only.

Need decide word count 560-840. Original text maybe ~830 excluding table. Let’s count. Paragraph 1 ~95. Gravity section ~130. small puddle ~50. Heat input ~180. callout ~45. Transfer ~190. Aluminum ~55. Uphill ~170. Downhill ~120. total ~1035 + table. Too long. Need cut to under 840. Keep all links. Links: Kemppi, Miller, weldzen aluminum, ESAB, weldzen all position. Facts: gravity, surface tension, arc force, small puddle, heat input equation behavior, transfer modes, aluminum half temp, uphill vs downhill. Need table maybe “Variable / What increases it / What reduces it” but truncated. We can reconstruct likely rows: Puddle size? Heat input? Travel speed? Need not invent? Table can be about variables. We can include a compact table with known relationships, but must keep same facts and maybe table was part. Since fragment truncated, table incomplete. We can include a table summarizing variables. But word budget. Need safe.

Need ensure links naturally. The internal links point weldzen.com, not weldgallery? Prompt says internal link targets on weldgallery.com but fragment has weldzen. Must keep all its links. We keep anchor text/URLs exactly. Need maybe not add other.

Need count words carefully. Let’s draft 700-780 words. HTML tags don’t count likely, headings count. We can aim ~720 words.

Need preserve heading exactly:

The Physics of Fighting Gravity: What the Puddle Is Actually Doing

Structure:
– intro ~75
– Gravity Drains, Surface Tension Holds ~130
– small puddle + note? ~50
– Heat Input ~150
– callout ~45
– Transfer Mode ~170 incl list
– Aluminum ~40
– Uphill vs Downhill ~160
– table ~30
Total ~810. Need careful.

Let’s write concise.

Need avoid em dashes overuse. The current has many. We can use commas/colons.

Potential table:

Variable What increases it What reduces it
Puddle size Higher voltage, slower travel, wider weave Lower voltage, faster travel, tighter weave
Heat input per inch More amperage or voltage, slower travel Less amperage or voltage, faster travel
Fluidity Spray transfer, long arc Short-circuit or pulsed transfer, tight arc

But “what increases it” for puddle size: higher voltage yes, slower travel yes, wider weave yes. For heat input: more amps/volts, slower travel. Good. This table adds maybe 35 words.

Need keep “same facts”. The fact “aluminum melts roughly half temperature of steel” is false-ish? Aluminum melts 660C, steel 1500C, so about 40%, roughly half. Keep.

Need make paragraph about transfer:
– short circuit keeps pool small, wire touches work, shorts dozens times per second, default out of position, deposits modest heat, freezes fast.
– spray fluid puddle, flat/horizontal, impossible vertical up.
– pulsed spray-like at lower avg current, thin/heat-sensitive, one droplet per cycle background current. Miller link.
Need not overdo.

“Kemppi’s overview…” link. “Miller’s pulsed MIG guide”. Internal weldzen links exact.

ESAB link fact: “positional progression changes both penetration profile and defect risk”. Keep.

Need word count. Let’s draft and count roughly.

I’ll compose final.

Draft:

…

Lay a plate flat and the puddle does half the work. Gravity presses molten metal into the joint, the arc stays short, and the weld freezes about where it landed. Tip the plate vertical and move uphill, and those forces flip. Gravity now pulls the pool down the face, while arc force and surface tension push it back into the groove and sidewalls. The bead is the visible score of that tug-of-war. Knowing which force is winning, second by second, separates a weld that looks like a caterpillar from one that carries load.

~90

Gravity Drains, Surface Tension Holds

A molten pool is held together by surface tension, the same force that makes water bead on a hood. On a vertical-up bead, gravity tries to drag the pool downward before the arc preheats the metal above and the weld freezes.

The arc helps. Arc force pushes metal into the groove and against the sidewalls, which is why a tight arc on an uphill run drives penetration instead of just sitting on top. Let the arc length grow and that force drops sharply. The puddle grows hotter and larger, and gravity wins. The pool sags, rolls over the solidified bead, and leaves convex, lumpy ripples plus undercut where the toes should have tied in.

This is why a smaller puddle is easier to control. Less molten metal means less mass for gravity to accelerate and less time before it freezes. Settings, weave width, and travel speed all serve one goal: keep the pool small, narrow, and freezing quickly.

Heat Input…

Heat input depends on voltage, amperage, and travel speed, and beginners ignore travel speed. Heat input per inch rises with voltage and current but falls as you move faster, because the same energy spreads over more length. That creates a trap. When an uphill bead sags, the instinct is to slow down and watch the puddle. Slowing adds heat where you already have too much, enlarges the pool, and gives gravity more time to drain it. Speeding up reduces heat per inch, but too much speed outruns fusion at the sidewalls. The usable window is narrow, which makes vertical-up slower and less forgiving than flat work even though it penetrates deeper on thicker steel.

~120

Callout.

Why Transfer…

The transfer mode sets the ceiling on puddle fluidity before you touch the dial:

  • Short-circuit transfer (dip) keeps the pool small because the wire touches the work, shorts, and clears dozens of times per second. It is the default for out-of-position work because it deposits modest heat and freezes fast. Kemppi’s overview… explains how each mode changes arc characteristics.
  • Spray transfer produces a highly fluid puddle of fine droplets. It is excellent flat, but uphill the liquid pool is nearly impossible to hold.
  • Pulsed transfer gives spray-like transfer at lower average current, which is why it appears on thin and heat-sensitive material. Miller’s pulsed MIG guide explains that each pulse delivers one droplet while background current keeps the puddle from building.

Aluminum is the classic test. It melts at roughly half the temperature of steel and conducts heat away fast, so a mild-steel setup will wash out an aluminum joint. Our Aluminum MIG Welding… covers how settings shift for non-ferrous work.

~150

Why Uphill…

Uphill progression keeps the arc on the leading edge of a small pool with cold metal above it, so more energy goes into the root instead of melting a large bath. confirms the trade-off: progression changes penetration profile and defect risk. Deeper penetration brings two risks. Undercut appears when the arc eats the sidewall faster than filler fills the toe, usually from too much voltage, a weave held too long at the edge, or a travel angle pointing sideways. Convexity appears when metal stacks in the center without wetting the toes, usually from too little heat or a weave too narrow for the joint.

Downhill reverses the physics. The arc pushes metal into a pool gravity is already spreading, so the bead is flatter, smoother, and faster to lay. The cost is dwell time. The arc passes quickly, preheats less, and penetration is shallower, raising the risk of lack of fusion on thicker sections. Vertical-down is a sheet-metal tool, not a structural default. For broader guidance on acceptable progression, see ….

…
Need total. Let’s estimate words:
Intro 86.
Gravity 45 + 78 + 48 =171. Actually first para 40, second 72, third 44 =156.
Heat 122.
Callout 44.
Transfer intro 12 + list 48+27+43=130 + aluminum 44 =186.
Uphill 112 + downhill 83 + table 35 =230.
Headings ~25.
Total 86+156+122+44+186+230+25=849. Slightly over 840. Need trim 20-40. Table words maybe 25. Hmm.

Let’s count precisely-ish. We can write shorter. Target 780.

Need include all links. Let’s trim intro, heat, transfer, uphill.

Maybe combine small puddle with gravity. Remove some.

Let’s write final around 760. Count by section.

I’ll use a word counter mentally. Better write text then count using rough. Since no tools supposed? We cannot use tools. We can count manually. But likely okay. Need between 560 and 840. 800 is fine. Original was maybe 900. Aim ~790.

Let’s produce with table. Ensure table doesn’t add too much.

Need table not “same” if truncated. It’s okay.

Let’s draft final exactly, count words per paragraph.

Heading 10? “The Physics of Fighting Gravity: What the Puddle Is Actually Doing” 10.

P1:
“Lay a plate flat and the puddle does half the work. Gravity presses molten metal into the joint, the arc stays short, and the weld freezes near where it landed. Tip the plate vertical and move uphill, and those forces flip. Gravity pulls the pool down the face while arc force and surface tension push it back into the groove and sidewalls. The bead is the visible score of that tug-of-war. Knowing which force is winning, second by second, separates a weld that looks like a caterpillar from one that carries load.”
Words: Lay1 a2 plate3 flat4 and5 the6 puddle7 does8 half9 the10 work11 Gravity12 presses13 molten14 metal15 into16 the17 joint18 the19 arc20 stays21 short22 and23 the24 weld25 freezes26 near27 where28 it29 landed30 Tip31 the32 plate33 vertical34 and35 move36 uphill37 and38 those39 forces40 flip41 Gravity42 pulls43 the44 pool45 down46 the47 face48 while49 arc50 force51 and52 surface53 tension54 push55 it56 back57 into58 the59 groove60 and61 sidewalls62 The63 bead64 is65 the66 visible67 score68 of69 that70 tug-of-war71 Knowing72 which73 force74 is75 winning76 second77 by78 second79 separates80 a81 weld82 that83 looks84 like85 a86 caterpillar87 from88 one89 that90 carries91 load92. 92.

Gravity heading 7. “Gravity Drains, Surface Tension Holds” 5? Gravity1 Drains2 Surface3 Tension4 Holds5.

P2: “A molten pool is held together by surface tension, the same force that makes water bead on a hood. On a vertical-up bead, gravity tries to drag the pool downward before the arc preheats the metal above and the weld freezes.”
A1 molten2 pool3 is4 held5 together6 by7 surface8 tension9 the10 same11 force12 that13 makes14 water15 bead16 on17 a18 hood19 On20 a21 vertical-up22 bead23 gravity24 tries25 to26 drag27 the28 pool29 downward30 before31 the32 arc33 preheats34 the35 metal36 above37 and38 the39 weld40 freezes41.

P3: “The arc helps. Arc force pushes metal into the groove and against the sidewalls, which is why a tight arc on an uphill run drives penetration instead of just sitting on top. Let the arc length grow and that force drops sharply. The puddle grows hotter and larger, and gravity wins. The pool sags, rolls over the solidified bead, and leaves convex, lumpy ripples plus undercut where the toes should have tied in.”
The1 arc2 helps3 Arc4 force5 pushes6 metal7 into8 the9 groove10 and11 against12 the13 sidewalls14 which15 is16 why17 a18 tight19 arc20 on21 an22 uphill23 run24 drives25 penetration26 instead27 of28 just29 sitting30 on31 top32 Let33 the34 arc35 length36 grow37 and38 that39 force40 drops41 sharply42 The43 puddle44 grows45 hotter46 and47 larger48 and49 gravity50 wins51 The52 pool53 sags54 rolls55 over56 the57 solidified58 bead59 and60 leaves61 convex62 lumpy63 ripples64 plus65 undercut66 where67 the68 toes69 should70 have71 tied72 in73.

P4: “This is why a smaller puddle is easier to control. Less molten metal means less mass for gravity to accelerate and less time before it freezes. Settings, weave width, and travel speed all serve one goal: keep the pool small, narrow, and freezing quickly.”
This1 is2 why3 a4 smaller5 puddle6 is7 easier8 to9 control10 Less11 molten12 metal13 means14 less15 mass16 for17 gravity18 to19 accelerate20 and21 less22 time23 before24 it25 freezes26 Settings27 weave28 width29 and30 travel31 speed32 all33 serve34 one35 goal36 keep37 the38 pool39 small40 narrow41 and42 freezing43 quickly44.

Running 92+5+41+73+44=255 + h2 10 =265.

Heat heading 8? “Heat Input: The Variable Behind Most Vertical-Up Problems” = Heat1 Input2 The3 Variable4 Behind5 Most6 Vertical-Up7 Problems8.

P: “Heat input depends on voltage, amperage, and travel speed, and beginners ignore travel speed. Heat input per inch rises with voltage and current but falls as you move faster, because the same energy spreads over more length. That creates a trap. When an uphill bead sags, the instinct is to slow down and watch the puddle. Slowing adds heat where you already have too much, enlarges the pool, and gives gravity more time to drain it. Speeding up reduces heat per inch, but too much speed outruns fusion at the sidewalls. The usable window is narrow, which makes vertical-up slower and less forgiving than flat work even though it penetrates deeper on thicker steel.”
Count:
Heat1 input2 depends3 on4 voltage5 amperage6 and7 travel8 speed9 and10 beginners11 ignore12 travel13 speed14 Heat15 input16 per17 inch18 rises19 with20 voltage21 and22 current23 but24 falls25 as26 you27 move28 faster29 because30 the31 same32 energy33 spreads34 over35 more36 length37 That38 creates39 a40 trap41 When42 an43 uphill44 bead45 sags46 the47 instinct48 is49 to50 slow51 down52 and53 watch54 the55 puddle56 Slowing57 adds58 heat59 where60 you61 already62 have63 too64 much65 enlarges66 the67 pool68 and69 gives70 gravity71 more72 time73 to74 drain75 it76 Speeding77 up78 reduces79 heat80 per81 inch82 but83 too84 much85 speed86 outruns87 fusion88 at89 the90 sidewalls91 The92 usable93 window94 is95 narrow96 which97 makes98 vertical-up99 slower100 and101 less102 forgiving103 than104 flat105 work106 even107 though108 it109 penetrates110 deeper111 on112 thicker113 steel114.
114.

Callout: “Quick Note: If the puddle sags, resist the urge to slow down. Check voltage and weave width first, then try a small increase in travel speed. Heat input per inch is what gravity responds to, not how carefully you move.”
Quick1 Note2 If3 the4 puddle5 sags6 resist7 the8 urge9 to10 slow11 down12 Check13 voltage14 and15 weave16 width17 first18 then19 try20 a21 small22 increase23 in24 travel25 speed26 Heat27 input28 per29 inch30 is31 what32 gravity33 responds34 to35 not36 how37 carefully38 you39 move40. 40.

Running 265+8+114+40=427.

Transfer heading: “Why Transfer Mode Decides How Much Heat You Get” 9? Why1 Transfer2 Mode3 Decides4 How5 Much6 Heat7 You8 Get9.

P intro 13.
Ul items:
LI1: “Short-circuit transfer (dip) keeps the pool small because the wire touches the work, shorts, and clears dozens of times per second. It is the default for out-of-position work because it deposits modest heat and freezes fast. Kemppi’s overview of MIG/MAG arc types explains how each mode changes arc characteristics.”
Count: Short-circuit1 transfer2 dip3 keeps4 the5 pool6 small7 because8 the9 wire10 touches11 the12 work13 shorts14 and15 clears16 dozens17 of18 times19 per20 second21 It22 is23 the24 default25 for26 out-of-position27 work28 because29 it30 deposits31 modest32 heat33 and34 freezes35 fast36 Kemppi’s37 overview38 of39 MIG/MAG40 arc41 types42 explains43 how44 each45 mode46 changes47 arc48 characteristics49. 49.

LI2: “Spray transfer produces a highly fluid puddle of fine droplets. It is excellent flat, but uphill the liquid pool is nearly impossible to hold.”
Spray1 transfer2 produces3 a4 highly5 fluid6 puddle7 of8 fine9 droplets10 It11 is12 excellent13 flat14 but15 uphill16 the17 liquid18 pool19 is20 nearly21 impossible22 to23 hold24. 24.

LI3: “Pulsed transfer gives spray-like transfer at lower average current, which is why it appears on thin and heat-sensitive material. Miller’s pulsed MIG guide explains that each pulse delivers one droplet while background current keeps the puddle from building.”
Pulsed1 transfer2 gives3 spray-like4 transfer5 at6 lower7 average8 current9 which10 is11 why12 it13 appears14 on15 thin16 and17 heat-sensitive18 material19 Miller’s20 pulsed21 MIG22 guide23 explains24 that25 each26 pulse27 delivers28 one29 droplet30 while31 background32 current33 keeps34 the35 puddle36 from37 building38. 38.

P aluminum: “Aluminum is the classic test. It melts at roughly half the temperature of steel and conducts heat away fast, so a mild-steel setup will wash out an aluminum joint. Our Aluminum MIG Welding: The Complete Practical Guide covers how settings shift for non-ferrous work.”
Aluminum1 is2 the3 classic4 test5 It6 melts7 at8 roughly9 half10 the11 temperature12 of13 steel14 and15 conducts16 heat17 away18 fast19 so20 a21 mild-steel22 setup23 will24 wash25 out26 an27 aluminum28 joint29 Our30 Aluminum31 MIG32 Welding33 The34 Complete35 Practical36 Guide37 covers38 how39 settings40 shift41 for42 non-ferrous43 work44. 44.

Section total 13+49+24+38+44+9=177. Running 604.

Uphill heading: “Why Uphill Penetrates Deeper, and Where It Goes Wrong” 8? Why1 Uphill2 Penetrates3 Deeper4 and5 Where6 It7 Goes8 Wrong9.

P: “Uphill progression keeps the arc on the leading edge of a small pool with cold metal above it, so more energy goes into the root instead of melting a large bath. ESAB’s MIG process parameter guide confirms the trade-off: progression changes penetration profile and defect risk. Deeper penetration brings two risks. Undercut appears when the arc eats the sidewall faster than filler fills the toe, usually from too much voltage, a weave held too long at the edge, or a travel angle pointing sideways. Convexity appears when metal stacks in the center without wetting the toes, usually from too little heat or a weave too narrow for the joint.”
Count: Uphill1 progression2 keeps3 the4 arc5 on6 the7 leading8 edge9 of10 a11 small12 pool13 with14 cold15 metal16 above17 it18 so19 more20 energy21 goes22 into23 the24 root25 instead26 of27 melting28 a29 large30 bath31 ESAB’s32 MIG33 process34 parameter35 guide36 confirms37 the38 trade-off39 progression40 changes41 penetration42 profile43 and44 defect45 risk46 Deeper47 penetration48 brings49 two50 risks51 Undercut52 appears53 when54 the55 arc56 eats57 the58 sidewall59 faster60 than61 filler62 fills63 the64 toe65 usually66 from67 too68 much69 voltage70 a71 weave72 held73 too74 long75 at76 the77 edge78 or79 a80 travel81 angle82 pointing83 sideways84 Convexity85 appears86 when87 metal88 stacks89 in90 the91 center92 without93 wetting94 the95 toes96 usually97 from98 too99 little100 heat101 or102 a103 weave104 too105 narrow106 for107 the108 joint109. 109.

P downhill: “Downhill reverses the physics. The arc pushes metal into a pool gravity is already spreading, so the bead is flatter, smoother, and faster to lay. The cost is dwell time. The arc passes quickly, preheats less, and penetration is shallower, raising the risk of lack of fusion on thicker sections. Vertical-down is a sheet-metal tool, not a structural default. For broader guidance on acceptable progression, see All Position Welding Rod Guide: Types, Uses & Tips.”
Downhill1 reverses2 the3 physics4 The5 arc6 pushes7 metal8 into9 a10 pool11 gravity12 is13 already14 spreading15 so16 the17 bead18 is19 flatter20 smoother21 and22 faster23 to24 lay25 The26 cost27 is28 dwell29 time30 The31 arc32 passes33 quickly34 preheats35 less36 and37 penetration38 is39 shallower40 raising41 the42 risk43 of44 lack45 of46 fusion47 on48 thicker49 sections50 Vertical-down51 is52 a53 sheet-metal54 tool55 not56 a57 structural58 default59 For60 broader61 guidance62 on63 acceptable64 progression65 see66 All67 Position68 Welding69 Rod70 Guide71 Types72 Uses73 &74 Tips75. 75.

Table:
Variable1 What2 increases3 it4 What5 reduces6 it7
Puddle size8 Higher9 voltage10 slower11 travel12 wider13 weave14 Lower15 voltage16 faster17 travel18 tighter19 weave20
Heat input per inch21 More22 amperage23 or24 voltage25 slower26 travel27 Less28 amperage29 or30 voltage31 faster32 travel33
Fluidity34 Spray35 transfer36 long37 arc38 Short-circuit39 or40 pulsed41 transfer42 tight43 arc44
44.

Subtotal 9+109+75+44=237. Running 841. Oops 841 plus maybe counts off. At upper bound 840. Need trim slightly. Also total 604 includes heading. Total 841. Need below 840, ideally ~790. Trim ~40-50.

Cut table rows maybe only two rows (25 words), saving 19. But need table maybe okay. Could omit table? Current fragment had table, but we can include compact. Need 560 min not issue. Aim 800.

Trim heat paragraph by 15 words, transfer by 10, intro by 10. Let’s adjust.

P1 reduce: “Lay a plate flat and the puddle does half the work. Gravity presses metal into the joint, the arc stays short, and the weld freezes near where it landed. Tip the plate vertical and move uphill, and those forces flip. Gravity pulls the pool down the face while arc force and surface tension push it back into the groove. The bead is the visible score of that tug-of-war. Knowing which force is winning, second by second, separates a weld that looks like a caterpillar from one that carries load.”
Removed “molten”, “and sidewalls” = 3. Need more.

Could remove “The bead is the visible score…” maybe keep? “That tug-of-war leaves a visible score.” Save 5.

Intro ~84.

Heat: “Heat input depends on voltage, amperage, and travel speed. Beginners ignore travel speed. Heat input per inch rises with voltage and current but falls as you move faster, because the same energy spreads over more length. That creates a trap. When an uphill bead sags, the instinct is to slow down and watch the puddle. Slowing adds heat where you already have too much, enlarges the pool, and gives gravity more time to drain it. Speeding up reduces heat per inch, but too much outruns fusion at the sidewalls. The usable window is narrow, which makes vertical-up slower and less forgiving than flat work, even though it penetrates deeper on thicker steel.”
Count roughly 110. Save 4.

Transfer: Aluminum paragraph “Aluminum is the classic test. It melts at roughly half the temperature of steel and conducts heat away fast, so a mild-steel setup will wash out an aluminum joint. Our Aluminum MIG Welding: The Complete Practical Guide covers how settings shift for non-ferrous work.” remove “classic test” maybe “Aluminum is the classic case” 1 less. Keep facts.

Could remove table entirely, but current fragment had table—user says same facts, not necessarily table. But table adds value. 44 words. If remove, total ~797. That seems safe and meets. But maybe losing table okay? Current fragment includes table truncated, but we need output full section; table likely expected. We can include a smaller 2-row table ~30 words, total 827. Hmm.

Let’s target 810 with 2-row table. Calculate accurately after trims.

Need ensure “all of its links” = 5. Yes.

Potential final length ~820. Fine. But need better than exact boundary because my counts approximate. Let’s aim 780. Remove table or make 3 columns 2 rows.

Let’s recalc with modifications.

P1 revised:
“Lay a plate flat and the puddle does half the work. Gravity presses metal into the joint, the arc stays short, and the weld freezes near where it landed. Tip the plate vertical and move uphill, and those forces flip. Gravity pulls the pool down the face while arc force and surface tension push it back into the groove. That tug-of-war leaves a visible score. Knowing which force is winning, second by second, separates a weld that looks like a caterpillar from one that carries load.”
Count: Lay1 a2 plate3 flat4 and5 the6 puddle7 does8 half9 the10 work11 Gravity12 presses13 metal14 into15 the16 joint17 the18 arc19 stays20 short21 and22 the23 weld24 freezes25 near26 where27 it28 landed29 Tip30 the31 plate32 vertical33 and34 move35 uphill36 and37 those38 forces39 flip40 Gravity41 pulls42 the43 pool44 down45 the46 face47 while48 arc49 force50 and51 surface52 tension53 push54 it55 back56 into57 the58 groove59 That60 tug-of-war61 leaves62 a63 visible64 score65 Knowing66 which67 force68 is69 winning70 second71 by72 second73 separates74 a75 weld76 that77 looks78 like79 a80 caterpillar81 from82 one83 that84 carries85 load86. 86.

Gravity P2 41, P3 73, P4 44 =158 + heading5.
Section = 86+158+5=249, plus h2 10 =259.

Heat heading 8 + paragraph revised:
“Heat input depends on voltage, amperage, and travel speed. Beginners ignore travel speed. Heat input per inch rises with voltage and current but falls as you move faster, because the same energy spreads over more length. That creates a trap. When an uphill bead sags, the instinct is to slow down and watch the puddle. Slowing adds heat where you already have too much, enlarges the pool, and gives gravity more time to drain it. Speeding up reduces heat per inch, but too much outruns fusion at the sidewalls. The usable window is narrow, making vertical-up slower and less forgiving than flat work even though it penetrates deeper on thicker steel.”
Count: Heat1 input2 depends3 on4 voltage5 amperage6 and7 travel8 speed9 Beginners10 ignore11 travel12 speed13 Heat14 input15 per16 inch17 rises18 with19 voltage20 and21 current22 but23 falls24 as25 you26 move27 faster28 because29 the30 same31 energy32 spreads33 over34 more35 length36 That37 creates38 a39 trap40 When41 an42 uphill43 bead44 sags45 the46 instinct47 is48 to49 slow50 down51 and52 watch53 the54 puddle55 Slowing56 adds57 heat58 where59 you60 already61 have62 too63 much64 enlarges65 the66 pool67 and68 gives69 gravity70 more71 time72 to73 drain74 it75 Speeding76 up77 reduces78 heat79 per80 inch81 but82 too83 much84 outruns85 fusion86 at87 the88 sidewalls89 The90 usable91 window92 is93 narrow94 making95 vertical-up96 slower97 and98 less99 forgiving100 than101 flat102 work103 even104 though105 it106 penetrates107 deeper108 on109 thicker110 steel111. 111.

Callout 40.

Transfer: heading 9 + intro 13 + list 49+24+38 + alum 44 =177.
Running: 259 + 8 +111+40 +177 =595. Wait 259+8=267, +111=378, +40=418, +177=595. Correct.

Uphill heading 9 + P 109 + P 75 + table 44 =237. Total 832. Still close. Need 800.

Trim 35 words:
– Table 2 rows ~25 instead of 44 saves 19.
– Transfer list 1 shorten save 6.
– Uphill paragraph shorten save 10.
Total ~798.

Make table 2 rows:
Variable / What increases it / What reduces it
Puddle size / Higher voltage, slower travel, wider weave / Lower voltage, faster travel, tighter weave
Heat input / More amps or volts, slower travel / Fewer amps or volts, faster travel
Words: Variable1 What2 increases3 it4 What5 reduces6 it7 Puddle8 size9 Higher10 voltage11 slower12 travel13 wider14 weave15 Lower16 voltage17 faster18 travel19 tighter20 weave21 Heat22 input23 More24 amps25 or26 volts27 slower28 travel29 Fewer30 amps31 or32 volts33 faster34 travel35. 35. Saves 9 only. Hmm.

Could make no table? Save 44, total 788. Is table required? Not necessarily. But current fragment had table likely “Variable | What increases it | What reduces it”. The output should be corrected fragment, same facts. Removing table is acceptable but perhaps section expects table. But word count hard. 832 under 840 technically, but my counts may be off by ~10. Better 790.

Let’s trim prose 40 and keep table 35 => 823. Need 15 more. Or remove table => 788. Best safe: keep table, trim 25 to 807.

Let’s trim:
P3 gravity: “The arc helps. Arc force pushes metal into the groove and sidewalls, so a tight arc on an uphill run drives penetration instead of sitting on top. Let the arc length grow and that force drops sharply. The puddle grows hotter and larger, and gravity wins. The pool sags, rolls over the solidified bead, and leaves convex, lumpy ripples plus undercut where the toes should have tied in.”
Removed “which is why” -> “so”, “just”. Save 4. 69.
P4: “A smaller puddle is easier to control. Less molten metal means less mass for gravity to accelerate and less time before it freezes. Settings, weave width, and travel speed all serve one goal: keep the pool small, narrow, and freezing quickly.” Removed “This is why” = 3. 41.
Saves 7.

Heat: remove “That creates a trap.” (4) and “even though it penetrates deeper…” maybe keep fact. Could say “The window is narrow, making vertical-up slower and less forgiving than flat work, though it penetrates deeper on thicker steel.” Save 7. Total 100.

Transfer:
LI1: “Short-circuit transfer (dip) keeps the pool small because the wire touches the work, shorts, and clears dozens of times per second. It is the default for out-of-position work since it deposits modest heat and freezes fast. Kemppi’s overview of MIG/MAG arc types explains how each mode changes the arc.” Save 3.
LI3: “Pulsed transfer gives spray-like transfer at lower average current, which is why it appears on thin and heat-sensitive material. Miller’s pulsed MIG guide explains that each pulse delivers one droplet while background current keeps the puddle from building.” Keep.
Alum: “Aluminum is the classic case. It melts at roughly half the temperature of steel and conducts heat away fast, so a mild-steel setup will wash out an aluminum joint. Our … covers how settings shift for non-ferrous work.” 41 save 3.
Total save 6.

Uphill:
P1: “Uphill progression keeps the arc on the leading edge of a small pool with cold metal above it, so more energy goes into the root instead of melting a large bath. ESAB’s MIG process parameter guide confirms the trade-off: progression changes penetration profile and defect risk. Deeper penetration brings two risks. Undercut appears when the arc eats the sidewall faster than filler fills the toe, usually from too much voltage, a weave held too long at the edge, or a travel angle pointing sideways. Convexity appears when metal stacks in the center without wetting the toes, usually from too little heat or a weave too narrow for the joint.” 109.
Could trim “with cold metal above it” essential. “so more energy goes into the root instead of melting a large bath.” Keep.
Maybe remove “Deeper penetration brings two risks.” (5) since next sentences explain. “Uphill’s deeper penetration brings two risks.” same.
“Uphill progression keeps the arc on the leading edge of a small pool with cold metal above it, so more energy goes into the root. ESAB… confirms the trade-off: progression changes penetration profile and defect risk. That deeper penetration brings two risks. Undercut appears when…” Remove “instead of melting a large bath” 7. Good.

Downhill: “Downhill reverses the physics. The arc pushes metal into a pool gravity is already spreading, so the bead is flatter, smoother, and faster to lay. The cost is dwell time. The arc passes quickly, preheats less, and penetration is shallower, raising the risk of lack of fusion on thicker sections. Vertical-down is a sheet-metal tool, not a structural default. For broader guidance, see All Position Welding Rod Guide: Types, Uses & Tips.” Remove “on acceptable progression” 3. 72.

New total: 832 -7 -11 -6 -10 -3 =795? Let’s see 832 -7 =825, heat save 11 =>814, transfer save 6 =>808, uphill save 10 =>798, downhill 3 =>795. Great. Table 35. Perfect.

Need recalc table maybe 35. Final 795.

But maybe initial total 832 included table 44, now table 35 =823, then prose saves 7+11+6+10+3=37 =>786. Fine.

Let’s write final at ~790. Need preserve all links. Need maybe “weldzen” titles. Exact anchor:
Aluminum MIG Welding: The Complete Practical Guide
All Position Welding Rod Guide: Types, Uses & Tips

External:
https://www.kemppi.com/en-AU/blogs/mig-mag-welding-arc-types
https://www.millerwelds.com/en-us/resources/knowledge-hub/mig-welding/how-to/guide-to-pulsed-mig-welding-in-manufacturing
https://esab.com/gb/eur_en/esab-university/articles/mig-welding-guide-process-parameters-and-best-practices/

Need maybe external target blank rel noopener. Current fragment didn’t have target, but rules say use. Add target=”_blank” rel=”noopener” for external. Internal links no target.

Need table HTML valid:

,

, th. Use compact.

Need check facts. “short circuits dozens of times per second” typical 20-200 Hz. Fine.
“spray limited to flat/horizontal” current says generally limited to flat and horizontal positions uphill, weird. We can say “excellent flat, but uphill…” safe.
“pulsed one droplet per cycle background current” Miller. Good.
“aluminum melts roughly half temperature steel” okay.
“uphill deeper penetration” accepted. “downhill shallower penetration and lack fusion risk” true.

Need “arc force pushes metal into groove” okay. “surface tension holds pool” yes.

Need maybe no “same facts” missing: “the arc preheats metal above it and freezes it in place” in original; new says before arc preheats above and weld freezes. okay.
“undercut where toes should tie in” yes.
“heat input falls as move faster” yes.
“if puddle sags, check voltage and weave width then speed” yes.
“travel speed narrow window” yes.
“transfer mode ceiling” yes.
“ESAB confirms positional progression changes…” yes.
“vertical down sheet metal tool not structural default” yes.

Need callout box. Good.

Now word count final might be under 840. Let’s produce.

One concern: “same heading” exact. Yes.

Need output only fragment, no commentary.

The Physics of Fighting Gravity: What the Puddle Is Actually Doing

Lay a plate flat and the puddle does half the work. Gravity presses metal into the joint, the arc stays short, and the weld freezes near where it landed. Tip the plate vertical and move uphill, and those forces flip. Gravity pulls the pool down the face while arc force and surface tension push it back into the groove. That tug-of-war leaves a visible score. Knowing which force is winning, second by second, separates a weld that looks like a caterpillar from one that carries load.

Gravity Drains, Surface Tension Holds

A molten pool is held together by surface tension, the same force that makes water bead on a hood. On a vertical-up bead, gravity tries to drag the pool downward before the arc preheats the metal above and the weld freezes.

The arc helps. Arc force pushes metal into the groove and sidewalls, so a tight arc on an uphill run drives penetration instead of sitting on top. Let the arc length grow and that force drops sharply. The puddle grows hotter and larger, and gravity wins. The pool sags, rolls over the solidified bead, and leaves convex, lumpy ripples plus undercut where the toes should have tied in.

A smaller puddle is easier to control. Less molten metal means less mass for gravity to accelerate and less time before it freezes. Settings, weave width, and travel speed all serve one goal: keep the pool small, narrow, and freezing quickly.

Heat Input: The Variable Behind Most Vertical-Up Problems

Heat input depends on voltage, amperage, and travel speed. Beginners ignore travel speed. Heat input per inch rises with voltage and current but falls as you move faster, because the same energy spreads over more length. When an uphill bead sags, the instinct is to slow down and watch the puddle. Slowing adds heat where you already have too much, enlarges the pool, and gives gravity more time to drain it. Speeding up reduces heat per inch, but too much outruns fusion at the sidewalls. The window is narrow, making vertical-up slower and less forgiving than flat work, though it penetrates deeper on thicker steel.

Quick Note: If the puddle sags, resist the urge to slow down. Check voltage and weave width first, then try a small increase in travel speed. Heat input per inch is what gravity responds to, not how carefully you move.

Why Transfer Mode Decides How Much Heat You Get

The transfer mode sets the ceiling on puddle fluidity before you touch the settings dial:

  • Short-circuit transfer (dip) keeps the weld pool small because the wire touches the work, shorts, and clears dozens of times per second. It is the default for out-of-position work since it deposits modest heat and freezes fast. Kemppi’s overview of MIG/MAG arc types explains how each mode changes the arc.
  • Spray transfer produces a highly fluid puddle of fine droplets. It is excellent flat, but uphill the liquid pool is nearly impossible to hold.
  • Pulsed transfer gives spray-like transfer at lower average current, which is why it appears on thin and heat-sensitive material. Miller’s pulsed MIG guide explains that each pulse delivers one droplet while background current keeps the puddle from building.

Aluminum is the classic case. It melts at roughly half the temperature of steel and conducts heat away fast, so a mild-steel setup will wash out an aluminum joint. Our Aluminum MIG Welding: The Complete Practical Guide covers how settings shift for non-ferrous work.

Why Uphill Penetrates Deeper, and Where It Goes Wrong

Uphill progression keeps the arc on the leading edge of a small pool with cold metal above it, so more energy goes into the root. ESAB’s MIG process parameter guide confirms the trade-off: progression changes penetration profile and defect risk. That deeper penetration brings two risks. Undercut appears when the arc eats the sidewall faster than filler fills the toe, usually from too much voltage, a weave held too long at the edge, or a travel angle pointing sideways. Convexity appears when metal stacks in the center without wetting the toes, usually from too little heat or a weave too narrow for the joint.

Downhill reverses the physics. The arc pushes metal into a pool gravity is already spreading, so the bead is flatter, smoother, and faster to lay. The cost is dwell time. The arc passes quickly, preheats less, and penetration is shallower, raising the risk of lack of fusion on thicker sections. Vertical-down is a sheet-metal tool, not a structural default. For broader guidance, see All Position Welding Rod Guide: Types, Uses & Tips.

Variable What increases it What reduces it
Puddle size Higher voltage, slower travel, wider weave Lower voltage, faster travel, tighter weave
Heat input More amps or volts, slower travel Fewer amps or volts, faster travel

Machine Setup: Voltage, Wire Speed, Gas, and Transfer Mode

Before you fight the puddle, set the machine so it stops fighting you. A surprising share of vertical-up problems that get blamed on shaky hands are actually setup problems: too much voltage, a stick-out that crept long, or a gas mix chosen for flat spray welding. Ten deliberate minutes at the machine saves an hour of grinding later.

How a Constant-Voltage Machine Really Behaves

Every MIG welder runs on a constant-voltage (CV) power source, and the two main controls do completely different jobs. Wire feed speed is your current control: feed more wire and the machine must burn it off faster, so amperage climbs; slow the feed and amperage drops. Voltage is your arc length control: more voltage gives a longer, hotter, wetter arc, while less voltage gives a short, stiff arc that freezes the puddle in place.

This is why CV machines suit uphill work so well. Because voltage holds steady while the wire feed sets current, a slight hesitation in your travel does not kill the arc, it simply lets the puddle fill a fraction more before you move on.

Memorise this rule from Miller’s MIG basics for mild steel: reduce voltage and amperage by roughly 10 to 15 percent from your flat settings before you start climbing. Two smaller adjustments matter just as much:

  • Inductance softens or sharpens the arc. Lower inductance gives a crisper, colder short circuit; higher inductance lets the puddle wet out and run smoother. Uphill, a touch more inductance usually helps the toes tie in without climbing.
  • Contact tip-to-work distance (stick-out) acts as a resistor in the circuit. Longer stick-out drops amperage, which sounds handy, but it also destabilises the arc and invites wire stubbing. Keep it close to the wire diameter, about 1/2 inch for most mild-steel work, and hold that distance constant as the gun travels upward.

Quick Note: Set your parameters on the machine, then forget the dials. Reading the display mid-weld is how arc length drifts. Adjust once, weld a test bead, adjust once more.

Choosing a Transfer Mode for Uphill Work

Transfer mode describes how droplets cross the arc, and it shapes heat input more than any single dial does.

  • Short-circuit (dip) transfer is the default for vertical-up. The wire touches the work, short-circuits, and melts off roughly 20 to 200 times per second, building a small, fast-freezing puddle. Kemppi puts short-arc parameters around 50 to 180 A and 15 to 23 V, and ESAB notes that dip transfer produces small weld pools with low heat input, which is exactly why it suits thin material and positional welding.
  • Globular transfer sits awkwardly in between. Kemppi describes it as a spattering arc type that should normally be avoided. If your bead throws coarse sparks and sits proud of the joint, you have probably drifted into globular without meaning to.
  • Spray transfer normally begins above roughly 24 V and 180 A in argon-rich gas. Uphill, that volume of molten metal is more than surface tension can hold, so it tends to sag or run out of the joint. Save spray for flat and horizontal fillets.
  • Pulsed MIG is the upgrade path. It delivers one droplet per pulse at lower average heat, keeping the puddle small while approaching spray-like penetration. It is genuinely useful on thin gauge and stainless vertical work, as covered in Miller’s guide to pulsed MIG.

Wire, Gas, and Polarity

Run DCEP (electrode positive) unless the wire manufacturer tells you otherwise. Reversed polarity gives the deepest penetration and the most stable arcing with solid wire.

For wire diameter, match the wire to the thinnest part of the joint, not the thickest:

Material thickness Wire diameter Typical use
24–18 ga (0.6–1.2 mm) 0.023 in Sheet, thin wall tubing, body panels
18–12 ga (1.2–2.7 mm) 0.030 in General fabrication, light repairs
1/8–1/2 in (3–12 mm) 0.035 in Structural plate, multi-pass vertical
3/8 in and heavier 0.045 in Heavy sections, high-deposition passes

ER70S-6 is the sensible mild-steel filler: its deoxidisers tolerateER70S-6 is the sensible mild-steel filler: its deoxidisers tolerate light mill scale and still run clean. For gas, a 75/25 argon-CO₂ mix is the standard short-circuit choice uphill; step up to 85/15 or 90/10 only if you move to pulse or spray. Mix options and cylinder handling are covered in this AllGas Welding Supply guide.

Thickness Wire Gas WFS (ipm) Volts
18 ga sheet 0.023 in 75/25 150–180 16–18
1/8 in 0.030 in 75/25 180–220 18–20
3/16–1/4 in 0.035 in 75/25 220–280 20–22
3/8 in+ 0.035–0.045 in 75/25 280–350 21–23

Note: Treat that chart as a starting point, not a setting. Run a bead on scrap, then change one variable at a time. A bead that sits proud with toes that refuse to wet in means too much wire for the voltage, not too little.

Step-by-Step Vertical-Up Technique: Stance, Angle, Arc Length, and Motion

Settings get the machine into the ballpark. Your body and gun angle decide whether the puddle stays where you put it. One idea runs through this whole section: keep the arc short, keep the puddle small, and let the wire feed push molten metal into the joint faster than gravity can drain it out.

Stance and Body Position

Set up so the gun travels naturally upward through the whole joint without you reaching. Feet shoulder width apart, joint roughly at chest height, elbows tucked in. If the end of the weld forces you to stretch, your arc length grows as you climb and the bead goes from tight to blown out in the last two inches. Move your feet and reposition instead of reaching.

Brace your forearm or the heel of your gun hand against the workpiece, a clamp, or a piece of scrap. A braced arm gives you a pivot point, and a pivot point gives you a consistent travel speed. Rest your off hand lightly on the panel for feedback, and keep your head off to the side so you can see both toes of the puddle, not just the arc glare.

Work Angle and Travel Angle

Work angle is the angle between the gun and the joint, viewed as a cross section. On a T-joint fillet, split the corner at roughly 45 degrees and bisect the angle of the joint. On a butt groove, keep the gun square to the joint, within about 10 degrees either side. Tilting the gun toward one plate loads that sidewall with heat and starves the other, which is exactly how one toe ends up undercut while the other looks fine.

Travel angle is the tilt along the direction of travel. Miller specifies a travel angle of 5 to 15 degrees from perpendicular for normal welding conditions in all positions, and warns that angles beyond 20 to 25 degrees cause more spatter, less penetration, and arc instability. For vertical-up work, that means a 5 to 15 degree drop from perpendicular. Tilt too far and the arc stops biting the root; the puddle just rides on top of cold metal.

Joint Work angle Travel angle Stick-out
Vertical fillet (T-joint) ~45°, bisecting the corner 5–15° drop from perpendicular 3/8–1/2 in
Vertical groove (butt) Square to the joint, ±10° 5–15° drop from perpendicular 3/8–1/2 in
Pulsed MIG, vertical-up ~45° 10–15° 5/8–3/4 in

Stick-Out and Arc Length

Short-circuit transfer wants a short arc, and stick-out is how you control it. Around 3/8 to 1/2 inch of wire past the nozzle keeps the arc tight and the puddle stiff. For pulsed MIG, Miller recommends 5/8 to 3/4 inch stick-out with roughly a 45 degree work angle and a 10 to 15 degree travel angle. Longer stick-out drops current, which sounds harmless until you realize it also kills penetration and lets the puddle sag. If the wire stubs into the plate, you are too close. If it hisses and sprays, you are too far.

Quick Note: Set your stick-out before the arc starts and hold it. Most “the machine is acting up” moments on vertical-up are really a gun drifting closer and farther from the plate as the welder climbs.

Starting the Puddle in the Right Place

Trigger in on the lower plate, a quarter inch below the joint root, not in the gap and not on the upper edge. Hold still for a beat until you see a distinct puddle form and wet into both sidewalls, then begin moving. Starting inside the gap dumps heat into air and gives you a burn-through hole; starting on top lets the puddle run downhill before it ever fuses the root.

Three Motions That Work

A slight weaving motion helps control the size, shape, and cooling of the puddle in vertical-up work, but the motion has to be small and deliberate. Three patterns cover almost everything:

  • Straight stringer with a toe pause. Run straight up, hold a fraction of a second at each toe, and move quickly across the center. Best for root passes and thin material where any weave invites sag.
  • Narrow triangle or Z-weave. Zig from toe to toe with a pause at each corner. Good for fill passes on fillets, and easy to keep consistent.
  • Cursive-e or C pattern. Sweep up one side, arc across the top, come down the other side. The classic vertical-up motion, and the one that gives the best sidewall fusion when you keep it tight.

Weave width should stay narrow, roughly three wire diameters or less. Wide weaves increase heat input, raise the risk of undercut and sag, and let the puddle cool unevenly across its width. If you feel like you are painting a fence, you are too wide.

Timing: Pause at the Toes, Hurry Across the Middle

The heat belongs at the sidewalls, not in the center. Pause at each toe long enough to see the puddle wet in and flatten, then cross the middle fast. The center of a vertical-up bead is where gravity does its damage, so the less time you spend there, the better. If the middle looks humped and glassy while the toes look cold, you are crossing too slowly.

Tie-Ins, Restarts, and Crater Filling

On a restart, strike the arc about half an inch up on the finished bead, let the puddle form, then walk it back down into the crater and climb out again. That re-melts the previous stop and blends the two passes instead of leaving a cold lap. At the end of a weld, pause and add a small circular motion to fill the crater before you release the trigger. A crater left unfilled is a crack waiting to happen, especially on thicker sections. If you are working through fit-up problems or switching to a different process for the root, the same short-arc discipline carries over; our aluminum MIG welding guide covers how the angles and travel speed change when the base metal conducts heat away five times faster.

Reading the Puddle as Feedback

Stop watching the arc and watch the puddle behind it. A puddle that is round and slightly leading the arc is happy. A puddle that elongates and runs ahead of the wire is too hot, so speed up or back off the wire feed. A puddle that sits flat and refuses to wet the toes is too cold. A keyhole opening up in a root pass means you are about to drop through and need to move. The arc noise tells you the transfer mode is stable, but the puddle shape tells you whether the weld will pass inspection.

Joint-by-Joint Playbook: Grooves, Fillets, Lap Joints, and Pipe

The technique in the previous section is only half the story. The other half is learning what each joint type actually asks of you, because a vertical groove in 3/8-inch plate and a vertical fillet on thin sheet share a position and very little else. This section translates the same short-arc, narrow-weave, toe-pausing approach into the four joint families you will meet most often: groove welds on plate, T-joint fillets, lap joints, and pipe.

Vertical Groove Welds on Plate: Root, Hot Pass, Fill, and Cap

Groove welds are where vertical-up earns its reputation. In structural work, vertical-up is the standard progression for notch-tough, structurally critical groove welds in many WPS documents, precisely because it drives the arc into the sidewalls instead of skating over them. The strategy in every groove pass is the same: establish fusion into both walls before you add any fill metal across the middle.

The Root Pass: Smallest Puddle of the Whole Job

Nothing else in vertical-up MIG punishes a big puddle faster than a root pass. Without backing, you are holding a keyhole open while gravity pulls the pool straight down into the gap, so run the smallest puddle you can establish and still fuse both lands. A land around 1/16 inch with a root gap in the 1/8-inch range is a common starting point, but gaps and lands are controlled by the procedure, not the mood of the moment.

With backing, life gets easier: the backing bar stops the pool from dropping through, so you can concentrate on sidewall fusion and a smooth pass profile. This is the configuration used in most structural qualification tests. AWS D1.1 welder qualification plates are commonly 3/8 inch (0.375 in) thick with backing, which in many agency records qualifies the welder from roughly 1/8 inch up to 3/4 inch — a reminder that thickness qualification is a range, not a single number. The AWS D1.1 structural welding code spells out those ranges and test assemblies in detail.

Hot Pass, Fill, and Cap: Sequencing That Keeps the Joint Flat

The hot pass goes in before the root has fully cooled. Its job is to burn out trapped gas and slag pockets and to smooth the root profile so later passes have something honest to sit on. Then fill passes build the joint up, alternating sides when weave width allows, and the cap finishes the surface.

Two habits separate a clean groove from a ropy one. First, clean between every pass — a wire brush and a light grinder touch on the toes, not a quick flick. Second, keep the cap flat. Ropy, highly convex caps usually come from too much wire feed and too little travel, stacking metal instead of tying into the sidewalls. Stringers or a very tight weave laid side by side give you a flatter cap than one wide, slow, wandering pass. Many procedures also limit weave width, often to roughly two to two-and-a-half times the electrode diameter, to cap heat input.

Quick Note: Consumable care matters even in MIG shops. Field notes from the Minnesota DOT welder qualification review specify drying E7018 electrodes at 450 to 500°F for at least two hours before use. If your shop also runs stick for repair or test plates, treat that oven schedule as mandatory, not optional.

The T-Joint Fillet: Your Everyday Vertical-Up Job

The 3F vertical fillet is the most common vertical-up case in fabrication and maintenance work, and it is forgiving enough to learn on. Set a 45-degree work angle so both plates get equal heat, drop the travel angle 5 to 15 degrees from perpendicular, and hold the toes for a beat while crossing the center fast. If you are used to flat fillets, the surprise is how much the puddle wants to slide toward the lower plate — your work angle correction is what stops the leg lengths from going lopsided. The all position welding rod guide is a useful cross-reference if you also run stick on the same joints.

Lap Joints and Thin Sheet

On sheet, burn-through replaces undercut as the main enemy. Point the wire slightly more into the thicker member, shorten the arc, and expect to reduce wire feed well below what the same thickness would take flat. On material under about 1/8 inch, many shops simply switch to vertical-down, which is faster and far less likely to drop a hole. If you are working aluminum, the same logic applies but the heat spreads faster, so read the aluminum MIG welding guide before you commit.

Pipe: The Vertical-Up Segment of a 5G and 6G Test

On pipe, only part of the circumference is true vertical-up. In a 5G test the pipe is horizontal and you weld around it; the section from roughly 4 o’clock to 8 o’clock — the bottom — is where uphill progression applies. In 6G the pipe sits at 45 degrees, and the vertical-up portion gets harder because your body position, not just the puddle, is fighting gravity. Keep the root keyhole small, pause at each sidewall, and tie in carefully at every restart; cold lap at a restart is one of the most common causes of a failed bend test.

Thickness Notes at a Glance

  • Under 1/8 inch: burn-through risk dominates. Short arc, low wire feed, consider vertical-down.
  • 1/8 to 1/4 inch: the sweet spot. Single-pass fillets and simple grooves behave predictably here.
  • 3/8 inch and up: beveling, multiple passes, and preheat on thick or restrained sections. Groove welds in thicker plate almost always require a bevel and several passes.
Joint Type Typical Approach Main Risk
Vertical groove, plate Root, hot pass, fill, flat cap Lack of sidewall fusion
T-joint fillet (3F) 45° work angle, tight weave Unequal legs, undercut
Lap joint, sheet Low wire feed, bias to thicker member Burn-through
Pipe 5G/6G uphill Small keyhole, careful restarts Cold lap at tie-ins

One last point that applies to all four: judge the finished joint by profile, not by decoration. A gently crowned, well-tied cap beats a showy ripple pattern every time, and the code does not grade aesthetics.

Troubleshooting Vertical-Up Defects: Undercut, Sag, Lack of Fusion, and More

Almost every ugly vertical-up bead traces back to one of three root causes: too much heat, too wide a weave, or a gun angle that stopped pointing where it should. The defect is just the signature. Read it, and you can make one deliberate change instead of guessing at the machine.

Before you touch a dial, check three things: where the defect sits (toe, center, root), the bead’s shape (flat, crowned, ropey, concave), and when it appeared (first inch, mid-pass, at a restart). A defect on one toe only is usually an angle problem. Both toes points to heat or weave width. Anything that shows up at a restart is technique, not settings.

Undercut at the Toes

Undercut is a groove melted into the base metal along one or both bead edges and left unfilled. It’s the most common vertical-up defect, driven by excessive voltage (which widens and softens the arc), travel too fast (you leave before the puddle wets the wall), or too short a pause at the sidewall.

The fix is rarely more wire. Drop voltage a volt or two, shorten the arc slightly, and hold each toe a clear half-beat longer than feels natural. If undercut follows one side of a fillet, check work angle first — a gun leaning away from a wall starves that wall of arc force and filler at the same time.

Convex, Ropy Beads and Puddle Sag or Rollover

A bead that sits up like rope, or a puddle that sags and rolls over the toes before freezing, means the molten pool is bigger than surface tension can hold. That’s a heat problem in disguise. Voltage too high, wire speed too high, travel too slow, or stick-out crept down to almost nothing all push heat in faster than the plate can shed it.

Correct it in this order: reduce wire feed speed slightly, then voltage, then increase travel speed. Keep the weave narrow — roughly three wire diameters wide is plenty for most vertical-up work. Widening the weave to flatten a ropy bead is the classic trap: it adds heat per unit length and makes the sag worse.

Quick Note: Chasing a defect with more wire feed almost always makes it worse. Extra wire means extra current, a bigger puddle, and more metal for gravity to drag down. When a vertical-up bead misbehaves, the first move is almost always less heat, not more filler.

Lack of Fusion at the Sidewalls

This one hides. The bead looks fine outside, then fails a bend test with unfused metal along one wall. In vertical-up MIG, lack of fusion links to travel speed and weave width far more than to amperage alone. Travel too fast and the arc never lingers long enough to melt into the sidewall; weave too wide and the arc spends most of its time over solidified metal instead of the joint faces.

The cure is a tighter weave, a genuine pause at each wall, and enough voltage to keep the arc wetting in. On a groove weld, watch the keyhole at the root — if it disappears, you’re moving too fast.

Burn-Through on Thin Metal

On sheet and thin-wall tube, vertical-up punishes you instantly. Burn-through opens as a hole from the inside out, usually right after the arc starts or at a slight gap. Reduce wire feed speed, step down to a smaller wire, shorten the stick-out, and consider vertical-down on the thinnest material. Small circles or a tight whip motion spread heat without stacking metal.

Porosity, Spatter, and Surface Signals

Porosity in MIG welds commonly traces to inadequate shielding gas coverage rather than to the metal itself. Check for a clogged or spattered nozzle, a draft from an open door or fan, gas flow set too low or too high (turbulence pulls in air), and contaminated base metal — mill scale, oil, paint, rust, or galvanized coating. Clean to bright metal and check flow at the cup, not the regulator.

Spatter rises sharply when transfer slides into the globular range, an arc type Kemppi describes as one that should normally be avoided. If your vertical-up welds suddenly sound harsh and throw large, irregular balls, you’ve drifted out of short-circuit transfer — usually too much voltage for the wire speed.

Wagon Tracks, Crater Cracks, and Arc Wandering

Wagon tracks are the deep parallel grooves left when the next pass fails to fuse into the shoulders of the one below. They come from poor interpass cleaning and from running the bead too far off center. Wire-brush and chip every pass. Crater cracks tear at the end of a bead as the puddle shrinks; fill the crater by pausing, adding a beat of wire, and breaking the arc back over the finished bead. Arc wandering is usually a connection issue — a poor work clamp, loose gun liner, or contact tip worn oval.

Quick Diagnostic Table

Defect Most likely cause First fix
Undercut at toes Voltage high, travel fast, short pause Lower voltage, pause longer
Convex or ropy bead Excess heat, slow travel, wide weave Reduce WFS, narrow the weave
Puddle sag or rollover Pool too large for surface tension Less voltage, faster travel
Lack of fusion Travel speed, weave width Tighten weave, hold at walls
Burn-through Too much heat for thickness Smaller wire, lower WFS
Porosity Gas coverage, draft, dirt Clean nozzle, shield arc, clean metal
Heavy spatter Globular transfer Lower voltage into short-circuit range
Wagon tracks Poor cleaning, off-center pass Clean between passes, recenter
Crater crack Unfilled crater Pause and fill before breaking arc

One last bit of housekeeping: verify your parameters against a reliable reference before blaming your hands. Miller’s guide to MIG basics for mild steel is a solid sanity check on voltage, wire speed, and stick-out. And if you’re troubleshooting a joint you don’t normally run, the all position welding rod guide helps explain how position changes deposition behavior.

Vertical Up vs Vertical Down vs Flat: An Honest Comparison

Ask three welders which progression is “best” and you will get three confident, contradictory answers. The truth is that flat, vertical-up, and vertical-down each solve a different problem. Choosing between them is not a matter of pride; it is a matter of thickness, joint access, code, and what the part has to survive after it cools.

Factor Flat (1G/1F) Vertical Up (3G/3F) Vertical Down (3G/3F)
Penetration Good, gravity-neutral Deepest of the three Shallow, root-focused
Travel speed Fastest Slowest Fast
Appearance Smooth, easy to shape Ropier, needs skill Flat and neat
Distortion Moderate Highest heat input Lowest heat input
Best thickness Any, with the right setup 1/4 in and up Sheet and thin wall
Main defect risk Lack of fusion if cold Undercut, sag, convexity Lack of fusion on thick steel
Code acceptance Widely accepted Commonly required Restricted or prohibited in many structural procedures

Where Vertical Down Earns Its Place

Vertical-down gets a bad reputation because it is misused. On thin material it is often the smarter choice. Miller notes that vertical-down helps on thin metal because faster travel speed means the arc penetrates less, and welders sometimes deliberately stand thin sheet vertically to avoid burn-through. The puddle chases the arc downward, heat stays concentrated at the surface, and the bead stays flat.

The trade-off appears as thickness climbs. On 3/8-inch plate, that shallow penetration and quick pass leave cold toes and possible lack of fusion at the root. Vertical-down is a finishing tool for thin sections, not a shortcut for structural plate.

Why Vertical Up Wins on Thick Steel

Miller states that vertical-up can provide better penetration on thicker materials, typically 1/4 inch or more. The arc force drives metal into the joint while the puddle sits above the arc, and that combination produces fusion you can trust under load. The cost is speed, heat input, and a narrower margin for error. If you are comparing consumables across positions, this all-position welding rod guide is a useful companion read.

Flat Is Easy, but It Is Not Always Available

Flat welding is the fastest and most forgiving position. Gravity holds the puddle where you put it, voltage tolerances widen, and appearance comes almost free. The catch is geometry. A trailer frame on a bench, a repair stand clamped at chest height, a pipe flange in place, or a joint overhead in a shop ceiling all remove the option. Rotating the work is always worth considering before you commit to a harder position.

Three Myths Worth Retiring

  • “Vertical-up is always stronger.” Not on 16-gauge sheet, where the extra heat burns through and the “stronger” weld becomes a hole.
  • “Vertical-down is always prohibited.” False. Many procedures accept it for thin material, and some specify it for root passes with backing.
  • “A vertical-up bead must look like a stack of dimes.” That look is a TIG habit, not a MIG requirement. A sound vertical-up MIG bead is often smoother and slightly convex, with consistent toe lines.

Quick Note: Progression is an essential variable. A procedure qualified uphill is not automatically valid downhill, so read the WPS before you assume either direction is approved.

What Codes Actually Say

Vertical progression is recorded on the WPS, which means direction is part of the qualification, not a personal preference. AWS D1.1 has historically included specific vertical-up welding requirements in its fabrication clauses, and the current edition is worth checking directly in the AWS D1.1 preview document. Some procedures qualify vertical-up only, and running downhill on that job is a nonconformance even if the weld looks perfect.

The practical takeaway: match the progression to the thickness and the paperwork. Thin sheet and root passes often favor down; plate 1/4 inch and heavier usually favor up; flat remains the default whenever the part can be turned.

Safety, Fume Control, and Getting Qualified: The Parts Most Guides Skip

Vertical-up welding has a dirty little secret that has nothing to do with technique: because you weld with your head close to the joint and the plume rises straight past your helmet, you tend to breathe more fume than you would on a flat bench. The same uphill direction that helps penetration also sends metal vapor and shielding-gas byproducts directly into your breathing zone. Get this part right before you chase a prettier bead.

What Is Actually in the Plume

Mild-steel MIG fume contains iron oxide, manganese, and trace metals. Stainless and chrome-containing consumables add another problem entirely. Chromium in the wire and base metal is converted to hexavalent chromium during welding, and OSHA describes Cr(VI) fume as highly toxic and carcinogenic. The permissible exposure limit is 5 micrograms per cubic meter as an 8-hour time-weighted average, with an action level of 2.5 micrograms per cubic meter. Cross that action level and your employer owes you exposure monitoring and medical surveillance.

Manganese matters too. OSHA notes that prolonged exposure to manganese fume can cause Parkinson’s-like symptoms, which is why neurology researchers keep studying welders specifically. For general welding fume, the air contaminants table in OSHA’s welding fact sheet lists 15 mg/m³ for total particulate and 5 mg/m³ for the respirable fraction. Those numbers look generous until you notice that individual metals carry far lower limits, and a combustible mixture of them does not average out to safety.

Ventilation, Position, and Respiratory Protection

The most common assumption in home shops is that an open garage door counts as ventilation. OSHA has stated plainly that welding outdoors or in open work spaces does not guarantee adequate ventilation, because fume can hang in your breathing zone regardless of the room’s total volume. Local exhaust ventilation is the fix that actually works. In a NIOSH-reviewed study, a bell-shaped hood placed 2 to 3 inches above the weld cut median hexavalent chromium concentrations by 68 percent. That is a bigger reduction than most respirator programs achieve in practice, simply because the hood catches fume before it reaches your face.

For work that cannot be captured at the source, a fitted half-mask with P100 particulate filters is the baseline; PAPR systems are more comfortable for long vertical passes where you cannot hold your breath and squint at the same time. Keep your head out of the plume, not just behind the shield.

PPE and Electrical Basics

  • Helmet shade: typically 10 to 12 for MIG at 100–200 amps, darker for higher current. If the puddle looks washed out, go up a shade.
  • Gloves and leathers: a vertically travelling gun drops spatter onto your forearms and lap. Wear sleeves and a leather apron.
  • Eyes: clear safety glasses under the helmet, always, for chipping slag and wire clipping.
  • Hearing: grinding and plasma cutting, not the arc itself, are the noise hazards.
  • Input power: check that your circuit and breaker match the machine’s rated input, and never weld in damp gloves or standing water.

The Qualification Path: WPS, PQR, and the 3G/3F Test

A welding procedure specification (WPS) is the written recipe: process, filler, gas, amperage range, voltage range, travel speed, joint detail, and position. The procedure qualification record (PQR) is the evidence that the recipe was tested and passed. Under AWS D1.1, position is an essential variable, so a welder qualified flat is not qualified vertical. A 3G test is a vertical-up groove weld on plate, usually with a backing strip or open root; a 3F test is the vertical fillet equivalent. Specimens typically face visual inspection plus bend testing or radiographic examination, and state programs publish their acceptance criteria in documents such as this field welder qualification review.

Quick Note: Practicing without a procedure teaches you to satisfy your eye, not a code. Print a WPS, tape the parameters to the table, and weld to the numbers. If you are still sorting out which electrodes and positions go together, this all position welding rod guide is a useful companion read.

If a test is on your horizon, spend a few weeks running the exact joint, position, and parameter range you will be tested on. Familiarity with one procedure beats scattered practice every time.

A Practice Plan That Actually Works, Common Mistakes, and Final Thoughts

Vertical-up MIG improves fastest when practice is deliberate rather than frequent. Burning through a spool in random passes builds bad habits. A four-week plan, one variable at a time, and an honest log will teach you more than months of guessing.

Four Weeks to a Repeatable Bead

Week Focus What Success Looks Like
1 Pad beads flat, then vertical stringers on scrap. No weave. Consistent bead width, steady stick-out, no sag.
2 T-joint fillets with a narrow triangle or Z-weave. Even legs, fused toes, no undercut.
3 Groove welds with backing: root, hot pass, fill, cap. Flat cap, clean sidewalls, good tie-ins.
4 Cut-and-bend tests plus controlled parameter sweeps. Defects found and traced to a specific setting.

Week one is about the machine and your hands, not pretty beads. Run flat pads until the sound and puddle feel familiar, then move to vertical stringers. Keep the weave out of it. Week two adds the joint that most shop work involves: the vertical fillet. Week three introduces multi-pass groove work, where cleaning between passes matters as much as the arc. Week four is where real learning happens, because you stop judging welds by appearance alone.

How to Run a Controlled Experiment

One-variable-at-a-time testing is the fastest way to learn what a machine and wire combination actually wants. Change voltage by half a volt, not three. Keep wire feed speed, gas flow, travel speed, joint preparation, and stick-out identical. Run three beads, label them, then compare. Consistent stick-out matters because it changes the effective current and arc stability, as Miller’s MIG basics guide explains.

A weld log recording voltage, wire feed speed, gas flow, travel speed, and joint prep turns random practice into repeatable results. Add stick-out and the date. When a bead finally runs clean, you will know exactly how to get back there. Cut-and-etch or bend tests reveal lack of fusion that visual inspection will miss, so use them before you declare a setting “good.” The bend-test requirements in AWS D1.1 show why welders qualify procedures with mechanical testing rather than looks.

Quick Note: Photograph each test bead beside its log entry. A picture of the bead, the settings, and the bend result is far more useful than a notebook full of vague comments like “ran okay.”

The Ten Mistakes That Keep Showing Up

  • Too much voltage: the arc flares, the puddle spreads, and undercut follows.
  • Weaving too wide: a wide puddle cools unevenly and loses sidewall fusion.
  • Wrong work or travel angle: the arc pushes metal away from the joint instead of into it.
  • Long stick-out: effective current drops and the arc becomes unstable.
  • Moving too fast at the toes: the edges never get enough heat to fuse.
  • Not cleaning between passes: slag, spatter, and oxide become tomorrow’s lack of fusion.
  • Poor ground: erratic current produces erratic beads and false conclusions.
  • Welding over rust: contamination creates porosity and spatter.
  • Starting the arc in the gap: the root blows through or fails to tie in.
  • Judging by sound alone: a steady crackle cannot show undercut or a cold toe.

Final Thoughts: Let the Puddle Lead

Vertical-up rewards patience and observation. The puddle tells you what to do next: if it rises too high, reduce heat or pause less; if the toes look sharp and glassy, slow down at the edges. Learn to read those signals instead of reacting with more wire or more voltage. If solo practice stalls, a structured aluminium welding course can sharpen your general technique and feedback habits, even though this guide focuses on steel.

Keep the weave narrow, the arc short, and the log open. Progress will feel slow for a week or two, then a bead will come out flat, fused, and evenly rippled. That is not luck. It is the result of controlled practice, and it is repeatable.

Frequently Asked Questions

Is MIG welding vertical up stronger than vertical down?

On thick material, generally yes. Travelling uphill lets the arc dig into the root before the puddle freezes, which produces deeper penetration and better fusion on plate roughly 1/4 inch and thicker. On thin sheet, vertical down is often the better choice because it limits heat input and burn-through, and the joint does not need deep penetration.

What voltage and wire speed should I use for vertical-up MIG?

Start about 10 to 15 percent below your flat-position voltage and wire speed, then tune on scrap of the same thickness and joint type. As a rough starting point for 1/8 inch mild steel with 0.030 inch ER70S-6 and 75/25 gas, many welders begin around 16 to 18 volts and 90 to 120 amps, but machine calibration varies widely.

Why does my vertical-up bead keep sagging or rolling over?

The puddle is too large for gravity and surface tension to hold. Reduce voltage and wire feed speed slightly, narrow your weave, increase travel speed a touch, and make sure your travel angle is not pushing too much wire into the pool. A long stick-out or a joint gap that is too wide will also cause sag.

Should I weave or run stringers going uphill?

Both work, but narrow is better. A slight weave or a small triangle pattern helps tie into both sidewalls and control cooling, while a straight stringer with a brief pause at each toe suits thin material and root passes. Wide, slow weaves add heat and are a leading cause of undercut and convexity.

Can I use spray transfer for vertical-up welding?

Usually no. Spray transfer produces a large, fluid puddle that runs out of a vertical joint, so it is normally limited to flat and horizontal work. Short-circuit (dip) transfer is the standard choice for vertical-up, and pulsed MIG is the practical upgrade when you want more deposition with controlled heat.

What is the difference between 3G and 3F?

3G is a vertical groove weld and 3F is a vertical fillet weld. The number is the position and the letter is the weld type: F for fillet, G for groove, with V indicating vertical. Both are common welder qualification tests, and both record whether the progression was uphill or downhill.

Do I need a smaller wire for vertical-up MIG?

It often helps. Smaller diameter wire lets you run lower current with a more stable short arc and a smaller puddle, which is easier to control uphill. Many welders use 0.030 inch for 1/8 to 1/4 inch steel and reserve 0.035 or 0.045 inch for heavier sections where deposition matters more than finesse.

How do I stop undercut at the edges of a vertical-up weld?

Pause slightly longer at each toe so the puddle can fill the melted groove, and reduce voltage or travel speed if the arc is digging too aggressively. If the undercut is only on one side, check your work angle, since an uneven angle puts more heat into one plate.

Is vertical-up MIG allowed by AWS D1.1?

Yes, and it is routinely qualified. AWS D1.1 treats vertical progression as a procedure variable, so a WPS qualified uphill covers uphill work, and a welder tested in 3G with uphill progression is qualified for that combination. Always work to the project WPS rather than to general rules of thumb.

How long does it take to get good at vertical-up MIG?

With deliberate practice on scrap, most hobbyists see usable beads within a few hours and consistent, code-quality-looking welds after several weeks of short, focused sessions. The limiting factor is usually puddle reading and body position rather than machine settings, so slow, observed practice beats long, mindless runs.

NH

Nicholas Harper

Editorial Team

Covering welding processes, equipment selection, and shop-floor practice for WeldZen — with every claim checked against the governing code.

Share this technical standard

Related Studies

Technical Dispatch

Get The Next Benchmark First

New field studies and calibration sheets, once a week. No spam.