Welding Waveforms Explained: Pulse Shape, Slope, Duty Cycle and Control Modes
Welding waveforms are how micro welding is programmed. In processes with millisecond cycle times, the shape of the current curve — rise, peak, hold, fall — decides whether a joint fuses cleanly, splashes, cracks, or never forms a nugget. This guide covers the common wave types, their parameters, how materials want the shape adjusted, and how to read a trace when a run drifts.
What a welding waveform is
The current-vs-time picture
A welding waveform plots the electrical quantity delivered to the joint — usually current, sometimes voltage — against time. Every controller parameter sits on that plot: peak is the maximum, pulse width is how long the curve holds it, slope is how sharply it rises and falls, duty cycle governs how much of each period sits above baseline. The waveform is the recipe the machine executes on every weld.
Why waveforms matter more in micro welding than in ordinary welding
A heavy weld spends seconds under the arc and small deviations smooth out. A micro weld lasts a few milliseconds — no time to correct a bad start, no thermal reservoir to buffer the joint. Whatever the waveform delivers in the first millisecond is roughly the finished nugget. Parameters that feel like tuning knobs at heavy scale become primary design variables.
Waveforms vs simple DC — how the picture changes
A DC weld is a rectangular pulse at one current for a fixed duration; fine for the simplest tab jobs. Shaped waveforms exist because certain materials fail on a rectangle: copper reflects heat into the tip, coated steel dumps zinc into the interface, aluminum cracks on fast solidification.
Common waveform types
Single pulse
One rectangular current burst: rise, hold, fall. On nickel, nickel-plated steel, and clean mild-steel tab work it is still the workhorse — fewer parameters, less drift. Limits appear when the material needs sequencing: a coating to burn off, a soft copper that would splash on a hard start, a hardenable steel that needs a slower cooling curve. Range of single-pulse jobs: resistance spot welders.
Multi-pulse (double / triple pulse)
A sequence of shorter pulses separated by cool-time gaps. Classic pattern on zinc-coated steel: pulse one blows off the coating, pulse two forms the nugget. Triple-pulse adds a temper pass on hardenable steels. Costs cycle time but recovers yield when a single pulse would leave splash or porosity.
Upslope (ramp-up)
A controlled rise from zero to peak instead of a hard step. Lets contact resistance stabilize before full current arrives and reduces initial spatter on soft or conductive metals. On copper alloys, upslope is often the difference between a repeatable process and a shift of erratic welds.
Downslope (ramp-down / hold)
The mirror: a controlled fall from peak. On alloys that crack on rapid solidification — aluminum, high-strength steels, heat-treatable copper — downslope keeps the pool solidifying through a gradient the metal can tolerate. A hard cut-off freezes the surface first and cracks the interior.
Preheat + main + temper
A three-phase profile treats the joint like a small heat-treatment. Preheat brings the interface to temperature without fusing. Main pulse forms the nugget. Temper runs at reduced current to slow cooling. Standard shape for high-hardenability steels where a cold quench produces a brittle nugget.
Waveform parameters and what each one changes
These are the parameters most modern controllers expose in software. Each is a lever on a specific piece of the physics — nugget size, splash, cracking, or heat balance. Reliable runs come from tuning two or three together, not from a single magic number.
| Parameter | Typical range | What it primarily changes | What goes wrong if it's off |
|---|---|---|---|
| Peak current | 2-15 kA (RSW); 5-300 A (pulse arc) | Nugget diameter and depth | Under = cold weld; over = expulsion |
| Pulse width | 1-20 ms (RSW); 0.5-100 ms (pulse arc) | Total energy; nugget growth time | Short = no fusion; long = burn-through |
| Slope | 1-500 A/ms rise; 1-200 A/ms fall | Splash on rise; cracking on fall | None = spatter and crater cracks |
| Duty cycle | 5-50% typical | Thermal recovery between welds | High = tip overheating; low = throughput loss |
| Hold time | 5-50 ms after pulse | Solidification under force | Short = pop-outs; long = tip sticking |
| Base current | 5-50 A (pulse arc) | Arc maintenance between peaks | Low = arc drops; high = between-pulse heat |
| Frequency | 50-500 Hz (AC RSW) | Line-synchronized energy delivery | Off-frequency = flicker on line-limited installs |
These are typical envelopes for micro-scale work. Actual settings depend on material, thickness, joint geometry, electrode class, and machine characteristics. Published tables are a starting point; the confirming step is a sample weld with the trace captured.
How different materials want the wave shaped
Nickel and nickel-plated steel
Single pulse or a short upslope covers most tab work. Moderate resistivity, no coating, no crack sensitivity in alloys typical for connector tabs and battery interconnects. Peak currents 3-6 kA for common thicknesses; pulse widths 3-10 ms. Complications appear only on foil under 0.1 mm, where pulse width has to shrink to avoid burn-through.
Copper and copper alloys
Hardest common resistance job because the base metal drains heat from the joint faster than it can build. Answer: a short, high-peak pulse with an upslope — enough current density to fuse a nugget before the surrounding metal absorbs the heat. Class 3 or Class 14 electrodes hold shape where Class 2 tips would mushroom quickly.
Coated steel (galvanized, tinned)
Almost always needs a multi-pulse profile. Pulse one vaporizes the coating and stabilizes contact resistance; pulse two forms the nugget. A single pulse on galvanized steel either burns through inconsistently or traps zinc as a porosity path. Cool-time between pulses is short — enough for volatiles to clear, not so long the interface goes cold.
Aluminum on resistance
Marginal to begin with — tenacious oxide, high thermal conductivity, narrow melting range. When attempted, downslope is critical: aluminum forms crater cracks on rapid solidification more reliably than almost any other production alloy. A hard cut-off produces a shrinkage cavity; a controlled downslope lets the pool solidify from the inside out.
Dissimilar-metal pairs (Cu-Ni, Cu-steel)
Dissimilar joints need the wave shape to compensate for asymmetric heat balance. The more conductive side pulls heat away faster, so the joint wants more energy on the less-conductive side. Asymmetric electrode force and a step or ramped peak can rebalance the heat.
Reading a waveform trace for defect diagnosis
The baseline: what a good trace looks like
A repeatable weld produces a repeatable trace. Rise clean, peak flat or matching the shaped-peak program, fall on the programmed slope. First diagnostic step: capture a dozen traces on a run producing acceptable welds and treat those as the reference.
Splash and expulsion signatures
Splash appears as a sudden dip or spike partway through the pulse. The interface has ejected molten metal, opening a temporary gap the controller either overcompensates for (spike) or accepts as increased resistance (dip). Cause: force-current mismatch — too much current for the applied force, a late force ramp, or electrode misalignment.
Cold weld signatures
A cold weld reaches peak but ends before the nugget grows. The tell is pulse width against a known-good reference, or an unusually flat voltage trace. Diagnostic: pull a section; if peak is on-target but nugget is undersized, extend pulse width or add hold.
Electrode-wear drift signatures
Electrode wear shows up gradually across a shift. As tip faces mushroom and cooling channels scale, contact resistance rises for the same setpoint and peak voltage drifts upward. Traces overlaid across a shift make this visible well before the peel test does. Mechanical side: electrode selection guide.
Coating-transfer signatures
First welds on coated steel produce a jagged trace as the coating (zinc, tin, brass) is displaced and partially transferred to the tip. After a short break-in it stabilizes. A trace that stays jagged all batch indicates coating not being cleared — usually first-pulse energy too low, or cleaning interval stretched too far.
Control modes: constant current vs constant voltage vs constant power
Constant current (the RSW default)
The controller holds programmed current regardless of contact-resistance drift during the pulse. Most predictable mode for spot welds on stable geometries: a fixed stack, a known electrode, a known force. Right tradeoff between simplicity and precision for most spot production.
Constant voltage
Holds voltage across the joint and lets current follow contact resistance. If resistance rises during the pulse (typical as a coating burns off), current falls. Mostly used in seam welding and roll-spot processes where voltage-based control is more stable on a moving contact.
Constant power
Holds the product of voltage and current constant, compensating for drift in either quantity. On thin-foil laminate work — battery tab welding, thin-film sensor bonding — thermal history has to stay tight across a batch, and constant power keeps energy per pulse consistent as tips wear. More demanding to commission but pays off on long runs.
How to pick the control mode for a line
Single-spot on fixed geometry defaults to constant current. Seam or roll-spot defaults to constant voltage or constant power depending on which quantity is more stable through the motion. Thin-foil work where thermal history matters most defaults to constant power. Broader machine framing: how to choose a micro welder.
Common misconceptions about waveforms
"More peak current = better weld"
Above the current a specific stack absorbs without expulsion, the joint splashes and the nugget shrinks. Maximum useful peak is a physical limit set by force, cooling, and material.
"Slope is a nicety, not a requirement"
On stable materials, slope is optional. On copper, coated steel, or thin-foil work, slope is what makes the process repeatable at all. Adding it after everything else is set is a frequent source of frustration on jobs that needed slope from the start.
"A waveform can be copied from a similar machine"
Waveforms are calibrated to a specific machine's electrode force, cooling capacity, transformer, and cable inductance. A profile that runs a joint on one machine is a starting point on another, not a drop-in.
Frequently asked questions
Why does slope matter so much on copper?
Copper conducts heat away from the joint faster than most metals. A hard step to peak drains heat into the tip instead of the joint, leaving splash or a burn. An upslope gives the interface a chance to build resistance and hold heat where it is needed.
How many pulses is "too many" in a multi-pulse profile?
Two or three pulses cover most production cases. Beyond that, added pulses usually mean the wrong profile choice. If a five-pulse profile is being considered, revisit force, cleaning, or electrode alloy first.
Can an electrode dressing problem be diagnosed from the waveform trace alone?
Trace drift is a strong indicator — rising secondary current at a target setpoint, or rising voltage at constant current, both point to face growth. Tip inspection confirms. Trace is early warning; peel test is confirmation. Mechanical framework: electrode selection guide.
Constant current vs constant power — when does the extra control mode pay off?
Constant power pays off when thermal history across a batch has to stay tight and both current and resistance are drifting. Thin-foil battery tab welding is the canonical case. If drift is minor and geometry stable, constant current is fine.
Do laser and pulse arc waveforms follow the same principles as resistance?
The framework — peak, width, slope, sequencing — carries across all three process families. Pulsed laser systems expose most parameters in software; continuous-wave modes shape fewer (laser micro welders). Pulse arc waveforms are shaped much like resistance, with the added variable of arc-length (pulse arc / micro TIG welders).