Welding Graph Explained: A Complete Guide to Reading Welding Charts, Curves, and Diagrams

Amperage, Voltage, and Wire Feed Speed Charts: Reading Setup Graphs by Process and Thickness

These are the graphs welders tape to the side of the machine: the setup-family parameter charts that turn “make it weld nicer” into real numbers. Each process’s chart controls something different. On a stick machine you dial amperage directly; on a MIG machine you dial wire feed speed and amperage follows. Read the chart without understanding that difference and you’ll chase settings all afternoon. For how the dials interact, our complete guide to MIG setup and settings walks through it step by step.

SMAW: Amperage by Electrode Diameter and Polarity

Stick charts are refreshingly simple. Rows are electrode diameter, columns give an amperage window, and polarity is fixed by the electrode classification. The classic rule of thumb is roughly 1 amp per 0.001 in of electrode diameter as a starting point. A 1/8 in E7018 lands in a wide window — about 90 to 160 A — and that width matters. The same rod at 90 A on a flat root pass and at 155 A on heavy plate are both correct readings. Polarity shifts the window too, so keep our breakdown of E7018 amperage, polarity and technique open while you dial in.

GMAW: Amps, Volts, and Wire Feed Speed Move Together

MIG charts are three-dimensional. Wire diameter, transfer mode, and material thickness all shift the numbers, and wire feed speed (WFS) is the real control — amp the machine up and the wire must feed faster to hold a stable arc length.

That first row is Hobart’s published ER70S-6 chart for 0.035 in wire on 1/8 in steel. The second shows why thickness changes everything: a 1/4 in fillet needs more than one pass at short-circuit, while spray transfer with 0.045 in wire and 90/10 gas lays that same fillet in a single pass. These are the numbers in the AWS GMAW/FCAW pocket handbook, worth cross-checking against any chart taped to your machine. Gun angle matters, too — a worn nozzle tip throws voltage readings off, so see our guide to contact tip selection and maintenance.

FCAW: Feed Rate and Deposition Rate Charts

Flux-cored charts add a column MIG charts don’t have: deposition rate, which is the whole point of FCAW. Self-shielded (FCAW-S) wires run on straight polarity, while gas-shielded (FCAW-G) wires run on reverse, so the feed-rate window changes with polarity as well as diameter. Check which one you’re holding before you trust the numbers — we explain the difference in DCEN vs DCEP for flux core. Higher current means faster melt-off, so the feed-rate line climbs steeply with amperage.

GTAW: Amperage, Tungsten, Cup, and Gas Flow

TIG is the most equipment-heavy chart because it lists tungsten size and colour, cup diameter, filler diameter, and gas flow alongside amperage. Two rules save time: size the tungsten to about 1 A per 0.001 in for AC work, and remember that joint geometry changes the required heat. CK Worldwide lists 125–145 A for a 1/8 in aluminium butt joint but 140–160 A for a fillet at the same thickness, because a T-joint pulls heat out in three directions instead of two. Aluminium is its own world, and our guide to TIG polarity: DCEN, DCEP and AC explains why balance settings shift these numbers again.

SAW: Current, Voltage, and Travel Speed

Submerged arc charts are the most forgiving and the most powerful, because the arc is buried under flux and the process is usually mechanised. Current sets penetration, voltage sets bead width, and travel speed ties them together. In automation, current and travel speed are locked into a recipe rather than dialled by hand, which is why robotic submerged arc welding is built entirely around parameter graphs.

Why the Same Wire Gives Different Numbers

Every manufacturer tests against its own electrode coating, gas mix, and machine output curve. A 0.035 in wire from one brand may list a 15 A wider window than another’s. Add wire chemistry, feed-roller condition, and cable length, and the scatter is real. The chart is a map of a region, not a street address.

Read setup charts by picking the right row for your diameter and thickness, taking the middle of the window, then adjusting for position, joint type, and how the arc sounds. The numbers get you close; the arc tells you the rest, and our guide to building real welding skills through deliberate practice is how you learn to hear the difference.

Heat Input and t8/5 Cooling Curves: The Graph That Decides Whether Your Weld Cracks

Every other graph in this guide describes what happens at the arc. This one describes what happens after the arc leaves — and that is where welds actually fail. Heat input and cooling time decide whether the metal next to your weld becomes tough, ductile ferrite or brittle, crack-prone martensite. Get them wrong, and a bead that looks perfect on Friday is a rejected part on Monday.

Heat Input: A One-Line Formula With Big Consequences

Heat input (HI) is the energy your arc delivers per unit length of weld, and it comes from a formula you can run on a phone calculator:

HI = η × V × I ÷ v

• η = process efficiency (heat actually transferred into the plate)
• V = arc voltage (volts)
• I = welding current (amps)
• v = travel speed (mm per second)

The result lands in joules per millimetre; divide by 1,000 for kJ/mm. Efficiency is where most people go wrong. Roughly, GMAW runs about 0.8, GTAW about 0.6 (a lot of arc energy radiates away), and SAW about 1.0 because the flux blanket traps nearly all of it. That difference matters: identical volts, amps and travel speed produce very different heat inputs depending on the process.

Worked example. You’re running GMAW spray transfer at 28 V, 250 A, travelling 5 mm/s.

HI = 0.8 × 28 × 250 ÷ 5 = 1,120 J/mm = 1.12 kJ/mm

Now change only the travel speed to 8 mm/s and you get 0.70 kJ/mm — a 38% drop with the dial settings untouched. That one variable is why “my numbers were right” so often isn’t the whole story.

t8/5 and R540: Two Ways to Say “How Fast Did It Cool?”

Heat input tells you energy per length. It doesn’t tell you how fast the weld cooled, because that also depends on how quickly heat escapes into the plate. The metallurgy world solves this with t8/5 — the time, in seconds, the weld and HAZ take to fall from 800 °C to 500 °C. That window is where austenite transforms, so it essentially stamps the final microstructure.

The alternative measure is the cooling rate at 540 °C, R540, quoted in °C/s. The two are related by a handy approximation:

R540 ≈ 300 ÷ t8/5

So t8/5 = 6 s corresponds to roughly 50 °C/s. Practical guidance flags cold cracking as most likely when heat input sits between about 0.8 and 2.0 kJ/mm, which typically means t8/5 of roughly 4 to 11 seconds. Below that band you’re quenching into martensite; far above it you’re cooking the HAZ into coarse grain and losing toughness.

Thick Plate vs Thin Plate: Why Geometry Changes the Graph

On thick plate, heat flows mainly downward into the mass of steel — the three-dimensional (3D) regime. Cooling is fast and largely set by heat input and preheat. On thin sheet, heat has nowhere to go but sideways along the plate, so it’s a two-dimensional (2D) problem, and cooling time scales with heat input squared divided by thickness squared. Doubling your heat input on thin sheet doesn’t just double t8/5 — it roughly quadruples it. That’s why formulas copied from a 25 mm plate job can be badly wrong on 3 mm sheet.

Preheat Is the Biggest Lever You Have

Cooling rate depends on how much hotter the weld is than its surroundings — the driving term is (T − Tp), where Tp is preheat or ambient temperature. That relationship is squared, which makes preheat disproportionately powerful. Going from a 25 °C shop to a 150 °C preheat cuts the thick-plate cooling rate by roughly 43%. No tweak to amps, volts or travel speed comes close to that.

Why Winter Jobs Crack

A fabricator I know had a run of brackets pass inspection all summer, then crack in service the following January. Nothing about the welding technique changed. Two things did. First, the steel arrived at 2 °C instead of 20 °C, which shortened t8/5 and pushed the HAZ toward martensite in exactly the way the graph predicts. Second, condensation and damp electrodes added hydrogen — and hydrogen plus hard martensite plus restraint is the classic recipe for delayed cold cracking, failures that surface hours or days later, long after the part has left the shop. The fix wasn’t exotic: preheat, dry consumables, and a heat input deliberately chosen for the middle of the t8/5 window rather than its floor.

If you weld higher-carbon or alloy steels, this graph family is the difference between a repair and a scrap bin. For the metallurgical modelling behind these curves, the Journal of Materials Science & Technology publishes modern SH-CCT prediction work, and NovaSolver’s thermal calculators let you estimate t8/5 and cooling rates without building a spreadsheet from scratch. Pair that with our preheat and hardened-steel guidance in Can You Weld Hardened Steel? What Works & What Doesn’t, and you’ll have both halves of the picture: the number, and what to do with it.

CCT Diagrams, HAZ Hardness, and Carbon Equivalent Charts Explained

Every other graph in this guide tells you what your machine and your arc are doing. This family tells you what your steel is doing after you walk away. If weld failures were crimes, the CCT diagram would be the security footage.

Continuous cooling transformation (CCT) diagrams are metallurgical maps. The horizontal axis is time, usually on a logarithmic scale. The vertical axis is temperature. Across the middle of the plot run a series of curved “C” or “S” shaped bands: the transformation-start and transformation-finish curves. Where those curves sit tells you how fast a given steel must cool before its crystal structure locks into bainite or martensite instead of the softer ferrite and pearlite you probably wanted.

Here is the part that matters to a welder: the phase transformation we care about in most carbon and low-alloy steels begins near 800 °C and finishes near 500 °C. That is exactly the window the t8/5 cooling time measures, and it explains why that one number carries so much weight in the previous section. Your cooling curve gets drawn onto the CCT diagram as a straight or gently curving line sloping down and to the right. Where it crosses the transformation bands decides what phase forms. Steep line, fast cooling, hard brittle structure. Shallow line, slow cooling, tough and ductile structure.

SH-CCT Diagrams and the Real Heat-Affected Zone

A standard CCT diagram assumes you heated a whole coupon uniformly. A weld doesn’t do that. It heats a narrow strip and lets it cool against cold parent metal. So engineers use simulated-HAZ CCT diagrams, or SH-CCT diagrams, produced by running thermal cycles that mimic an actual weld. Alongside the transformation curves, SH-CCT plots print a hardness value for each cooling rate. Read across from your t8/5, drop to the hardness scale, and you get a prediction of HAZ hardness before you strike a single arc.

The reason hardness is the number that matters comes down to hydrogen. Peak HAZ hardness above roughly 350 HV is the usual trigger for hydrogen-induced cold cracking, the delayed failure that shows up days after welding. In sour service, where sulfide stress corrosion cracking is a concern, specifications tighten that ceiling to about 248 HV. Cracked welds rarely announce themselves on the day you finish.

Carbon Equivalent: The Chemistry Behind the Curves

Why does one steel crack at 350 HV while another shrugs it off? Chemistry. Carbon equivalent (CE) folds the hardening contribution of carbon, manganese, chromium, molybdenum and other alloying elements into a single number. Raise CE and the Hmax-versus-t8/5 curve climbs and shifts left, meaning the same cooling rate now produces a harder, more crack-prone HAZ. High-CE steel needs more preheat and often more heat input to stay safe.

AWS D1.1 handles this without demanding a CCT diagram on site. Its Annex I hardness-control approach is applicable to steels with carbon above 0.11%, and it suppresses cracking by tying energy input and preheat together as a matched pair. You don’t get to raise the amps without also raising the preheat. The standard treats them as one decision, because metallurgically they are.

Modern Shortcuts: Machine Learning and Public Databases

Producing a fresh SH-CCT diagram used to take weeks of dilatometry testing. Now it takes seconds. A hybrid machine learning model combining multilayer perceptron, k-nearest neighbours and random forest predicted SH-CCT phase transformation temperatures and hardness from chemistry and cooling rate with reported classification accuracy up to 100% on its test data.

If you want to see what a real diagram looks like without buying one, Japan’s NIMS maintains a public weld CCT diagram database you can search by steel grade and welding condition. It is a genuinely useful reference when you are quoting a job in an unfamiliar material and want a sanity check before committing to a procedure.

One caution. Predicted diagrams are only as good as the input chemistry, and mill certificates report ladle analysis, not the exact heat under your hood. Treat any CCT plot as a strong indicator, then verify with hardness testing on a real coupon. A single portable hardness reading on the HAZ will tell you more than ten beautifully rendered curves — and it takes about thirty seconds.

The takeaway for the shop floor is simple. When a weld cracks in a hard, high-carbon-equivalent steel, the graph that explains it almost never lives in your machine manual. It lives in the metallurgy, and it was decided by how fast you let that HAZ cool.

Duty Cycle and Machine Output Graphs: What the Curve Says About Your Power Source

Your machine’s data plate carries a line most of us skim past: 60% duty cycle at 200 A. That single figure is a curve compressed into a caption, and knowing how it was drawn tells you far more about your power source than the amps dial ever will. Output graphs are the maps that keep machines alive and welders out of trouble.

Duty Cycle: A Percentage That Hides a Curve

Duty cycle is the percentage of a 10-minute period a machine can weld at a given output before thermal protection shuts the arc down. A 60% rating at 200 A means 6 minutes of arc time and 4 minutes of cooling out of every 10-minute window. Push past that and the fan keeps spinning while the arc dies — the machine is protecting its transformer and rectifier, not misbehaving.

Because output and duty cycle trade against each other, manufacturers plot them as a sloping curve: generous duty cycle at low amperage, tightening as current climbs. The rated output is what the machine holds continuously under standard test conditions. Maximum output is the short-duration peak. Reading both tells you whether a production job fits the box you bought.

Output Curves, the OCV Line, and Your Safety Margin

Three lines usually share that graph. The V-A output curve maps achievable amps and volts. The open circuit voltage (OCV) line marks the voltage waiting at the electrode before the arc strikes. The duty cycle curve rides along the same axes. Output itself may be AC, DC, or both, and may be constant current, constant voltage, or switchable between the two — which is why one capability chart can look so different from the next.

OCV matters twice over. It determines how easily the arc initiates and restarts, especially with cellulosic electrodes and long cable runs. It is also a hard electrical safety limit. OSHA caps maximum open circuit voltage depending on the type of current and how the machine is used, with familiar ceilings around 80 V for AC and 100 V for DC in manual arc welding. A high-OCV machine that strikes beautifully also puts more voltage on the stinger while you stand in damp boots. Setting up that machine safely comes down to layout and grounding as much as settings — see our guide to planning a safe welding shop. The AWS power source chapter preview is a useful free walkthrough of how these curves are generated and classified.

Pulsed Waveforms and AC Shapes: Graphs That Move

Some output graphs plot current against time rather than voltage against current, and that’s where things get interesting. Pulsed processes — which emerged from the 1970s onward alongside fast-response solid-state power sources — swing between a high peak current and a lower background current many times a second. On the graph, frequency is how tightly the pulses pack together, peak current drives the penetration spike, and background current keeps the arc lit between them. That waveform explains why pulse tames thin sheet and out-of-position work without changing your average amperage.

AC output gets the same treatment. A conventional sine wave crosses zero gradually, so the arc weakens and reignites every half cycle. A square wave transitions almost vertically, holding arc stability through the crossover and letting you adjust balance — shifting the cleaning-versus-penetration split that aluminium work depends on. Our explainer on TIG polarity, DCEN, DCEP, and AC shows how that balance changes what the arc actually does at the puddle.

Read the curve before you trust the dial. If you’re still matching a machine to your work, start with our beginner’s wire welder guide, then check the duty cycle curve at the amperage you’ll actually run.

Using Welding Graphs on the Shop Floor: WPS, PQR, and Parameter Windows

Out on the shop floor, the graph stops being theory. It shows up taped to a column or clipped into a job packet as a Welding Procedure Specification, and that single sheet decides whether two welders on the same joint produce the same bead. The catch is that most welders read a WPS like a recipe with one magic number. It isn’t. It’s a map with borders, and those borders are where the real story lives.

A WPS Is a Box, Not a Dot

A WPS lists a parameter window, never a single setting. That window is itself a kind of graph: a box drawn around allowable amperage, voltage, wire feed speed (WFS), travel speed, and contact-tip-to-work distance (CTWD). Stay inside the box and the weld is qualified by the procedure. Wander outside and you’re no longer welding to the spec, no matter how good the bead looks. The procedure qualification record (PQR) is the test that proved the box works; the WPS is simply the box written down for the floor. If that distinction still feels slippery, our explainer on building real welding skills through deliberate practice walks through how procedures and hands-on technique reinforce each other.

Some shops present FCAW and GMAW parameter data graphically for exactly this reason. Welders find it far easier to see that CTWD and voltage must rise as wire feed speed rises when the relationship is drawn as a curve rather than buried in a column of numbers.

How to Read a Graph-Style WPS

Here’s the sequence a welder or supervisor follows on a chart-style procedure:

1. Pick your electrode or wire diameter.
2. Start from the mid-range wire feed speed, not the top or bottom edge.
3. Move horizontally to the diameter line.
4. Read the voltage and the prescribed contact-tip-to-work distance.
5. Confirm the amperage falls inside the listed window.

CTWD is the variable most welders treat as a suggestion. It isn’t. As WFS increases, CTWD must increase to keep the arc stable and the current honest. Hold the gun too far from the work and you lose current, widen the arc, and risk lack of fusion. Hold it too close and current climbs, spatter jumps, and the tip overheats. That’s why a serious contact tip selection and maintenance routine matters as much as the chart itself.

Why the Window Has Legal-Edges

The ranges on a WPS aren’t decorative. AWS D1.1 prequalified variables require requalification for changes beyond 10% in amperage or WFS and 15% in voltage. Cross those lines and you owe the project a new procedure qualification. That’s also why factory data sheets need translation: a manufacturer’s table showing a 100% swing in wire feed speed must be narrowed to a usable window before it can drive a controlled production procedure. AWS has helped here by publishing Standard Welding Procedure Specifications (SWPSs) for common material and process combinations since 1990, giving shops a vetted starting point rather than a blank page. You can preview the structure in the AWS B2.1 SWPS preview.

Two Welders, One WPS, Two Different Beads

Picture a real scenario. Two welders run the same WPS on a fillet joint. The first welder works near the bottom of the amperage and WFS window; the second sits near the top. Both are technically “compliant,” but the beads look nothing alike. The low-end weld is cold and crowned; the high-end weld is flat, wide, and slightly undercut. The fix wasn’t rework or a new welding process. It was a narrower, graph-style WPS that collapsed the range to a tight mid-band, so both welders landed on the same curve. A window is a permission slip, not a plan.

Read the window, pick the center, adjust for position and thickness, then hold it steady. That’s how a graph turns into a weld that passes inspection the first time.

Digital Welding Graphs: Simulators, Data Logging, and AI-Assisted Charts

Twenty years ago, reading a welding graph meant squinting at a laminated curve taped to the side of a machine. Today the graph reads itself — and then emails you a copy. Inverters and welding management systems now sample current, voltage, wire feed speed, and travel speed several times a second, which means the entire weld can be plotted after the fact instead of estimated from a dial setting. That shift changes the job from setting a number to interpreting a data trail.

The upside is enormous. The risk is trusting a pretty curve more than the steel in front of you.

Arc Data Logging: Plotting the Whole Weld, Not the Dial

On a logged machine, every arc-on second produces a waveform: current spiking and settling, voltage riding up and down with arc length, travel speed drifting whenever the welder hesitates. Play that back and you can see things a visual inspection never reveals — an operator who ramps voltage 2 V high midway up a vertical joint, a wire feeder that stutters during a long pass, a travel speed that climbs as the weld gets easier to reach. Shops running robotic submerged arc welding automation lean on this heavily, because a robot’s graph should look identical on weld 1 and weld 400.

The practical takeaway: once your process is logged, your parameters stop being a guess and start being a record. That record is what lets you trace a defect back to a specific pass on a specific day.

Heat-Input and Cooling-Rate Calculators

Digital heat-input calculators take the arithmetic from earlier in this guide and extend it into thermal modelling. Most are built on the Rosenthal analytical solution for a moving point heat source, developed in the 1930s and 1940s and still the backbone of weld pool size and cooling-rate estimation. Feed in voltage, amperage, travel speed, thickness, and preheat, and the calculator returns heat input, t8/5, cooling rate, and an estimated HAZ width.

Good calculators switch between thick-plate (3D) and thin-plate (2D) behaviour using transition thicknesses, and that switch matters more than most welders realise. SEW 088 places the thick-to-thin transition at roughly 15.5 mm for 20 °C preheat, moving to about 17.4 mm once you preheat to 150 °C. Get the regime wrong and your cooling estimate can swing wildly, because thin-plate behaviour scales with heat input squared over thickness squared.

Simulators and Machine Learning CCT Predictions

Welding simulators have quietly become serious training tools. Trainees run an arc against a screen that responds to arc length, work angle, and travel speed, building the eye-hand habits described in our guide to welding practice before they burn a single expensive coupon.

On the metallurgy side, machine learning is now estimating simulated-HAZ CCT diagrams and HAZ hardness directly from chemical composition and cooling rate, cutting reliance on slow, costly dilatometry experiments. The research published in the Journal of Materials Science & Technology shows how far these predictions have come. Verified public data, such as the NIMS weld CCT diagram database, gives those models something honest to train against.

Where Digital Graphs Get It Wrong

Every analytical cooling model makes simplifying assumptions: constant thermal properties, no surface heat loss, and no latent heat of transformation. Compare a raw Rosenthal prediction with the experimentally corrected SEW 088 guideline and thick plate typically diverges by about 15%, while thin plate can be off by roughly 40%. Multi-pass work adds another trap — the interpass temperature should be treated as the initial temperature for the next pass, which many quick calculators ignore.

How to Use Digital Graphs Without Getting Burned

• Use calculators for trends and screening, not final acceptance.
• Confirm with hardness testing and a procedure qualification before production.
• Treat logged parameters as one input alongside visual and NDT results.
• Re-run the model whenever thickness, preheat, or position changes.

Digital graphs are maps with better ink, not a replacement for the compass in your hand. Use them to see the terrain earlier — then let the coupon, the hardness file, and the weld itself have the final word.

Common Mistakes, Practical Tips, and Final Checklist for Reading a Welding Graph

After ten sections of curves, windows and cooling times, here is the honest truth: almost nobody loses a weld because they couldn’t read a graph. They lose it because they read one number from a graph, set the machine to that number, and walked away. This final section is the closing checklist — the mistakes to stop making, the sequence to follow before you strike an arc, and the safety rules that matter once a chart convinces you to weld hotter and longer.

The Seven Mistakes That Cost Welders Money

• Treating a chart’s number as the target. A table entry is the middle of a window, not a bullseye. Land mid-range, then let the puddle tell you which way to move.
• Using flat-position amps for vertical or overhead work. Out-of-position welding sheds heat differently, so most procedures drop amperage roughly 10 to 15 percent and often reduce wire feed speed too.
• Ignoring travel speed in the heat input calculation. HI = η·V·I/v. Two welders at identical amps and volts can differ by 50 percent in heat input purely because of how fast they move.
• Applying the thick-plate formula to thin sheet. Thin plate behaves as a 2D heat sink, so cooling scales with heat input squared over thickness squared. Those numbers are not interchangeable.
• Forgetting that preheat and interpass temperature move the whole curve. A change from 25 °C ambient to 150 °C preheat can cut thick-plate cooling by more than 40 percent. It is the single biggest lever you own.
• Reading amperage off a CV machine. On GMAW and FCAW, wire feed speed is the real control. Voltage sets arc length; WFS sets current. Chasing a dial that doesn’t do what you think is wasted effort.
• Assuming one maker’s table transfers to another’s consumable. Flux composition, coating thickness and even spool winding shift the numbers. Use the filler manufacturer’s own data.

Your Pre-Arc Checklist

1. Identify the process and the machine characteristic. Is your power source CC (drooping, for SMAW and GTAW) or CV (flat, for GMAW and FCAW)? Everything downstream depends on this.
2. Pick the right graph family. Electrical, setup, thermal or verification — match the chart to the question you’re actually asking.
3. Find the window, not the point. Note the high and low limits, then start mid-range.
4. Adjust for position and thickness. Drop amps or WFS for vertical and overhead; check whether you’re in thin-plate or thick-plate territory.
5. Calculate heat input and t8/5. Jot the number down, even roughly. It tells you whether martensite is likely in the heat-affected zone.
6. Set preheat and cap your interpass temperature. Preheat slows cooling; excessive interpass heat coarsens the grain. Both limits exist for a reason.
7. Test on scrap of the same thickness and grade. There is no substitute for a real bead on real steel. Building that habit is the core of welding practice.
8. Record what actually worked. Your own logged settings become the most useful graph in the shop.

If you’re new to wire processes, wrapping the graph-reading habit around a solid setup routine makes it far easier. Our complete MIG welding at home guide pairs well with this checklist.

Safety: A Hotter Weld Means More Fume

Every graph that pushes you toward higher current, longer arc time and slower travel also pushes you toward more fume. Welding fume carries metals such as chromium, nickel, manganese, cadmium and lead, plus gases including ozone, nitrogen dioxide and carbon monoxide. Stainless steel is the sharp edge here: the chromium it contains converts to toxic hexavalent chromium, Cr(VI), in the arc. OSHA limits Cr(VI) exposure to 5 µg/m³ as an 8-hour time-weighted average under 29 CFR 1910.1026 and 1926.1126.

• Keep local exhaust ventilation close to the plume source, not across the shop.
• When welding outdoors, position yourself upwind and let the breeze carry fume away.
• Never let a graph justify welding through a cloud. Stop, ventilate, resume.

Graphs Are Maps. Experience Is the Compass.

Here’s the encouraging part. Every curve in this guide was drawn by someone who once stood exactly where you’re standing, watching a bead misbehave and guessing. A graph won’t make the decision for you. It simply narrows the guesswork, tells you which direction is dangerous, and shortens the distance between a cold lumpy bead and one that holds. Read the chart, set your window, run the test bead, and then trust your hands. The map shows the route. You still drive.

Frequently Asked Questions

What is a welding graph in simple terms?

A welding graph is any plotted relationship between welding variables — most often current and voltage, but also wire feed speed vs amperage, heat input vs travel speed, or cooling time vs HAZ hardness. It shows how changing one setting shifts the others, so you can predict arc behaviour and weld quality instead of guessing. The most fundamental welding graph is the volt-ampere output characteristic curve of the power source.

What is the difference between a constant current and constant voltage welding graph?

A constant current (CC) graph has a steep, drooping curve, so a big change in arc length produces only a small change in current — roughly ±8 A for a ±5 V change at 150 A. A constant voltage (CV) graph is flat, so the same arc-length change produces a much larger current swing of around ±40 A. CC machines suit stick, TIG and SAW; CV machines suit MIG and flux-core, where the arc is self-regulating.

How do I use a welding amperage chart for a 1/8 inch electrode?

Start with the one-amp-per-thousandth rule, which gives about 125 A for a 0.125 in rod, then check the manufacturer’s published range for your exact electrode class. A 1/8 in E7018 typically runs 90–160 A on DC+, while an E6010 of the same size runs about 70–130 A. Reduce the setting by 10–15% for vertical and overhead positions and always confirm with a test bead.

What is t8/5 and why does it appear on welding cooling graphs?

t8/5 is the time in seconds that the weld and heat-affected zone take to cool from 800 °C to 500 °C — the window where most of the phase transformation happens in carbon and low-alloy steel. A short t8/5 means fast cooling and hard, crack-prone martensite; a long t8/5 means slower cooling and softer, tougher HAZ. It is controlled mainly by heat input, plate thickness and preheat, and the approximate relationship R540 ≈ 300 ÷ t8/5 links it to the 540 °C cooling rate.

Can I just copy the numbers from a welding chart?

No — a chart gives you a starting window, not a prescription. Published ranges differ between manufacturers and even between editions of the same manufacturer’s tables, and they assume particular joint types, positions and gas mixes. Use the chart to pick a starting point inside the window, then run a test bead on scrap of the same material and thickness, inspect penetration and bead profile, and adjust from there.

Why did my weld crack when the amperage chart said my setting was fine?

Amperage alone does not control cracking — heat input does, and heat input depends on voltage, current and travel speed together. If you welded faster than the procedure assumed, heat input dropped, t8/5 shortened, and the HAZ hardened above roughly 350 HV, which combined with hydrogen can cause delayed cold cracking. Carbon equivalent, preheat, interpass temperature and moisture all shift that graph too.

What is a CCT diagram and do I need one?

A continuous cooling transformation (CCT) diagram plots how steel transforms as it cools at different rates, showing which microstructure — ferrite, pearlite, bainite or martensite — you will get, along with HAZ hardness. Welders do not read them daily, but engineers use them to set heat input and preheat limits for high-strength or high-carbon steels. Modern machine learning models can now estimate SH-CCT diagrams from chemistry and cooling rate.

Do welding graphs apply to MIG and flux-core the same way they do to stick?

Not exactly. On CC machines used for stick and TIG, amperage is the controlling preset and voltage is measured; on CV machines used for MIG and flux-core, voltage and wire feed speed are the presets and amperage is a result. That is why a MIG parameter graph usually plots wire feed speed against voltage, with amperage as a readout rather than an input.

Are there safety limits connected to welding graphs?

Yes. OSHA limits maximum open circuit voltage on arc welding power sources, which is a graph parameter, and fume generation rises with current and duty cycle, so hotter and longer welding means more exposure. OSHA’s permissible exposure limit for hexavalent chromium is 5 µg/m³ as an 8-hour time-weighted average, and welding fume contains manganese, nickel and other metals that require ventilation or respiratory protection.

Which welding graph should a beginner learn first?

Start with the amperage-by-diameter chart for the process you actually use, then learn the volt-ampere curve for your machine (CC or CV). Next, understand heat input as voltage times current divided by travel speed, then t8/5 cooling, then duty cycle. Those five cover the vast majority of day-to-day decisions on a fabrication floor.