Plasma & Oxy-Fuel Cutting: The 30,000 K Arc, the 5.6× Melt Energy, and the 1.5:1 Hole Rule Behind Every Plate Quote

Engineering guide to plasma and oxy-fuel cutting: arc physics, specific energy, oxygen purity, kerf, ISO 9013, dross, hole rules, HAZ and cost per metre.

·

Plasma & Oxy-Fuel Cutting: The 30,000 K Arc, the 5.6× Melt Energy, and the 1.5:1 Hole Rule Behind Every Plate Quote

A fabrication shop quoted six brackets in 15 mm S355, each with four 18 mm holes. The plasma table was a 200 A oxygen-plasma machine, well maintained, with a working torch height control. The programmer nested the parts, set the cut speed from the chart for 15 mm — 3,440 mm/min on a 300 A-class machine, so a little less on this one — and let the holes run at the same feed as the outside contour.

The parts came off the table with 1.2 mm of taper in every hole, a hard bead of dross along the bottom edge, and a top diameter 0.3 mm over nominal. The holes were 15.9 mm at the bottom and 18.3 mm at the top. An M16 bolt would not pass; an M18 would rattle. The shop slowed the machine by 10 %, which is the instinct, and the dross changed from a hard scalloped bead to a soft bubbly weld bead. Nothing got better.

Three separate errors had stacked up. An 18 mm hole in 15 mm plate is a 1.2:1 hole, and a conventional plasma torch is not a 1.2:1 tool — the practical floor is 1.5:1, and even that needs the hole feed cut to 60 % of contour speed. The nozzle had about 0.3 mm of orifice wear, so the kerf was wider than the CAM offset assumed. And the shop's oxygen supply was running at 99.2 % purity, which is a 7.5 % speed penalty before the first arc struck.

None of those errors is visible on a drawing. All of them are predictable from arithmetic. That is what this guide is about: the energy, gas, kerf and tolerance numbers that decide whether a plate gets cut cleanly on a plasma table, burned under an oxy-fuel torch, or sent to the grinder.

1. Two ways to separate steel

Plasma and oxy-fuel are both thermal cutting processes, and they are almost never interchangeable. They separate metal by different mechanisms.

Plasma cutting melts the metal and blows it out. A pilot arc is struck between an electrode and a nozzle, then transferred to the workpiece. The nozzle constricts the arc into a narrow column of ionised gas at up to 30,000 K, with energy densities up to 2 \times 10^6\ \text{W/cm}^2 and a jet velocity up to three times the speed of sound. The metal in the kerf melts, and the high-velocity gas ejects it. There is no chemical requirement on the workpiece: any electrically conductive material can be cut, including stainless, aluminium, copper and brass.

Oxy-fuel cutting burns the metal. A fuel gas flame heats the steel to its ignition temperature — roughly 1,050 °C, some 450 °C below the melting point — and then a jet of pure oxygen oxidises the iron. The reaction releases enough heat to raise the next layer to ignition temperature, so the cut is self-sustaining once started. The oxide melts at a lower temperature than the steel and is blown out by the oxygen stream. This works on low-carbon and low-alloy steel; stainless and aluminium form refractory oxides that do not melt at the cut temperature, so they cannot be cut this way without iron powder injection.

The practical consequences follow directly:

· Plasma · Oxy-fuel

Separation mechanism · Melt and eject · Oxidise and eject

Heat source · Electric arc · Fuel flame + iron oxidation

Peak temperature · up to 30,000 K · flame 2,800–3,160 °C; cut face ~1,500 °C

Materials · all conductive metals · carbon/low-alloy steel only

Thickness range · 0.5–150 mm (typical 1–50 mm) · 3–300 mm (up to 1,000 mm with large tips)

Kerf width · 2–6.5 mm · 2–5 mm

HAZ width · 2–5 mm · 4–10 mm

Typical ISO 9013 range · 2–4 · 3–5

Dominant operating cost · machine hour + electricity · machine hour + oxygen

Best at · thin-to-medium plate, holes, contours · thick plate, straight cuts, multi-torch

The economics reverse around 50–60 mm, and Section 11 works that crossover in rupees. Below it, plasma is faster and cleaner on almost every part.

2. The energy bill: 7.6 J/mm³ to melt, 43–86 J/mm³ to cut

The first number in any thermal cutting calculation is the enthalpy required to take steel from room temperature to a molten, superheated liquid. For steel it is:

That is the thermodynamic floor: 7.6 joules for every cubic millimetre of steel removed, assuming every joule goes into the kerf and nothing is lost to the bulk plate, the gas, or the nozzle.

Real plasma cutting consumes far more. The specific cutting energy is:

where P is arc power in watts, t the plate thickness in mm, k the kerf width in mm, and v the cutting speed in mm/s. Take a 200 A machine at 160 V cutting 12 mm steel at 2.5 m/min with a 1.5 mm kerf:

That is 5.6 times the melting enthalpy. The melt efficiency is therefore:

At 25 mm on a 300 A, 66.5 kW machine running 1,950 mm/min with a 1.8 mm kerf, the specific energy is 45.5 J/mm³ and the efficiency 16.7 %. At 50 mm, where the same machine is down to 560 mm/min and a 2 mm kerf, specific energy rises to 71.2 J/mm³ and efficiency falls to 10.7 %. Thick plate is not just slower; it is thermodynamically worse, because a larger fraction of the arc's heat is conducted sideways into the plate instead of removing the kerf.

Oxy-fuel's energy accounting is completely different. The cutting heat is the heat of formation of iron oxide:

Per kilogram of iron, that is 272/0.05585 \approx 4.87\ \text{MJ/kg}, or, at 7.85 g/cm³, about 38 J/mm³ of steel removed. On a 25 mm cut with a 2 mm kerf at 430 mm/min, the chemical heat release is roughly 1,900 J per millimetre of cut. The preheat flame supplies only 28–29 % of the total heat; the rest comes from burning the plate itself.

This is why the two processes fail differently. Plasma is limited by electrical power and arc stability. Oxy-fuel is limited by oxygen delivery and the rate at which the oxidation front can propagate through the plate.

3. Oxy-fuel: a chemical reaction with a speed limit

The stoichiometry of oxy-fuel cutting is unforgiving. One kilogram of iron needs:

of oxygen at STP, and real cuts use considerably more. At 25 mm and 430 mm/min, a 2 mm kerf removes about 169 g of steel per minute, which requires 34 L/min of oxygen stoichiometrically. A typical 25–40 mm nozzle flows 60 L/min. Utilisation is therefore about 56 % — roughly half the oxygen leaves the kerf unreacted. That excess is not waste in the usual sense: it provides the momentum that clears the oxide from the kerf, and starving it produces slag and a stalled cut.

The purity tax: one point of oxygen, 25 % of speed

Oxygen purity is the single most under-checked variable in flame cutting. The published threshold is 99.5 %; below about 95 % the cut degenerates into a melt-and-wash action, and below about 80 % steel cannot be cut at all. The penalty for small deviations is steep:

where x is the oxygen purity in percent. A drop of one point — 99.5 % to 98.5 % — cuts speed by about 25 % and raises oxygen consumption by about 25 %. Going from 99.5 % to 99.8 % buys only 6–10 % more speed, which is rarely worth the supply premium. The practical target is IS 309 Grade 2 (99.5 % minimum), and the practical failure is a cylinder bank or a liquid dewar that has drifted to 99.0–99.2 % without anyone noticing.

The mechanism is a stagnant boundary layer of nitrogen and argon at the burning face. Oxygen has to diffuse through it to reach the iron, and the reaction rate falls accordingly. It is not a flame-temperature problem; the preheat flame can look perfectly neutral while the cut runs 20 % slow.

Fuel gas and tip selection

Acetylene gives the hottest flame at about 3,160 °C in oxygen, and the fastest pierce. Propane runs around 2,800 °C, needs 4.3 volumes of oxygen per volume of fuel instead of 1.2, but is cheaper per metre and produces good cuts in thick plate. Natural gas sits between them. The choice matters most for piercing, not for steady-state cutting: once the cut is established, the oxygen jet, not the flame, does the work.

Oxy-fuel cut charts are expressed in tip size, cutting-oxygen pressure, and speed. A representative acetylene chart:

Plate thickness · Cutting oxygen · Speed · Kerf

5 mm · 3.5–7.0 bar · 580–735 mm/min · 1.0 mm

10 mm · 5.6–7.0 bar · 480–660 mm/min · 1.3 mm

25 mm · 5.6–8.0 bar · 380–510 mm/min · 1.5 mm

50 mm · 5.6–8.0 bar · 305–405 mm/min · 1.5 mm

100 mm · 5.6–8.0 bar · 230–305 mm/min · 2.0 mm

150 mm · 5.6–8.0 bar · 200–280 mm/min · 2.3 mm

250 mm · 5.6–8.0 bar · 100–150 mm/min · 4.3 mm

Two numbers stand out. First, speed falls by a factor of five between 5 mm and 250 mm while thickness rises fiftyfold — thick plate is not proportionally slower, it is disproportionately slower. Second, the kerf widens from 1 mm to 4.3 mm, which means kerf compensation must change with thickness, not just with material.

Piercing is the other limit. A plasma torch can pierce 32 mm steel in about 1.5 seconds. An oxy-fuel torch takes about 20 seconds to pierce 32 mm, because it has to preheat a spot to ignition temperature before opening the cutting oxygen. Above about 100 mm, piercing is impractical altogether and the cut must start from the plate edge. On a part with 60 holes, that difference is 23 minutes of pierce time on the gas torch against 90 seconds on plasma — which is why hole-heavy plate goes to plasma even when the plate is thick enough that oxy-fuel would cut it faster.

4. Plasma: constriction, supersonic gas, and why thickness is expensive

A plasma torch is a controlled explosion. The arc is struck between a tungsten or hafnium electrode and a copper nozzle, and the nozzle orifice constricts it. The constriction raises the current density: 200 A through a 1.5 mm orifice is about 113 A/mm². The gas — oxygen, air, nitrogen, argon-hydrogen, or F5 (5 % hydrogen in nitrogen) — is heated by the arc, dissociates and ionises, and expands out of the nozzle at supersonic velocity. On the workpiece the ions recombine and release their energy, which is why the heat transfer is so intense.

Arc power is simply:

with arc voltage typically 120–170 V and current from 30 A to 1,000 A. A 200 A cut at 160 V is 32 kW; a 300 A XPR-class machine at 222 V is 66.5 kW. That power is spread over a kerf that is only 1.4–3 mm wide, which is what produces the narrow HAZ and the fast cut.

The speed penalty with thickness is worse than linear. On a 300 A machine cutting mild steel with oxygen:

Thickness · Speed · Specific energy · Melt efficiency

15 mm · 3,440 mm/min · 43.0 J/mm³ · 17.7 %

20 mm · 2,550 mm/min · 43.5 J/mm³ · 17.5 %

25 mm · 1,950 mm/min · 45.5 J/mm³ · 16.7 %

30 mm · 1,530 mm/min · 48.3 J/mm³ · 15.7 %

40 mm · 940 mm/min · 53.1 J/mm³ · 14.3 %

50 mm · 560 mm/min · 71.2 J/mm³ · 10.7 %

60 mm · 385 mm/min · 86.4 J/mm³ · 8.8 %

Between 15 mm and 60 mm, thickness quadruples and speed falls by a factor of 8.9. The extra loss is conduction into the plate and a kerf that widens as the arc searches for a stable attachment. This is also why the same machine that cuts 25 mm at 1,950 mm/min is near its severance limit at 80 mm: the process is not simply slower, it is running out of energy density.

The cut height matters as much as the power. Plasma cut charts assume a torch standoff of about 1.5 mm, held by arc voltage control. Raise the torch and the arc voltage rises, the arc spreads, the kerf widens, and the cut edge bevels. Lower it and the nozzle risks contact and double-arcing. Pierce height is normally 2.5 times the cut height, with a pierce delay long enough for the arc to punch through — 0.2–1.0 s depending on thickness. On plate above about 40–45 mm, even a 300 A torch should edge-start rather than pierce.

5. ISO 9013: the only cut-quality language worth putting on a drawing

Most plasma disputes are not about whether the part is "good"; they are about which of two unstated standards applies. ISO 9013:2017 is the standard that resolves this. It classifies thermal cuts — plasma, laser and oxy-fuel — by two measurements:

Both scale with material thickness a in millimetres. The limits for the four main ranges are:

Plasma work commonly lands in Range 2–4: high-definition systems and thin plate hold 2–3, conventional systems and thicker plate drift into 4, and anything worse is Range 5. The Range 5 perpendicularity limit used by several plasma shops is:

At 10 mm, that gives the following:

Range · u · Rz_5 · Equivalent ISO 286 grade · Typical process

1 · 0.08 mm · 16 µm · IT9–IT10 · fibre laser

2 · 0.22 mm · 48 µm · IT10–IT11 · laser, best HD plasma

3 · 0.50 mm · 82 µm · IT11–IT12 · HD plasma

4 · 1.00 mm · 128 µm · IT12–IT13 · conventional plasma

5 · 1.55 mm · — · IT14–IT16 · oxy-fuel, rough plasma

The equivalence is approximate, but it is useful for pricing: a plasma cut at 10 mm is an IT12–IT13 feature, not an IT9 feature. If a drawing says "±0.2 mm" on a 10 mm plasma part without a finish operation, it is asking for something the process cannot hold on a consistent basis.

Taper, kerf and the worn nozzle

Bevel angle and perpendicularity are the same measurement in different units:

A conventional plasma cut bevels 3–5°; high-definition systems hold 0–3°, and the best hold about 1°. At 15 mm, 3° is 0.79 mm of radial taper, 5° is 1.31 mm. The bracket in the opening, with 1.2 mm of taper, was cutting at about 4.6° — inside ISO Range 5, outside Range 4. That is exactly the quality a conventional torch produces when it is asked to cut a 1.2:1 hole.

Kerf compensation is the other half of the tolerance problem. CAM offsets the toolpath by half the kerf — 0.7 mm for a 1.4 mm kerf, 1.5 mm for a 3 mm kerf. But the kerf is not constant. A nozzle orifice wears as it erodes, and the kerf can widen by up to about 0.4 mm over the life of a consumable set. The radial error is half of that, 0.2 mm, so a 10.00 mm hole becomes 10.40 mm in diameter with a worn nozzle, while an outside profile shrinks by 0.40 mm across. If the drawing tolerance is ±0.2 mm, the nozzle is the tolerance.

6. Dross: three species, one speed window

Dross is re-solidified metal that was not ejected from the kerf. It is the most common plasma defect and the least understood, because there are three distinct types with opposite causes.

Low-speed dross forms when the torch moves too slowly. The arc expands to maintain contact with the kerf walls, the kerf widens, and the high-velocity core of the jet no longer reaches the molten metal at the bottom. The result is a thick, bubbly, weld-bead-like accumulation that scrapes off easily. Remedies: increase speed in 125 mm/min steps, raise the standoff slightly, reduce amperage, or use a smaller nozzle.

High-speed dross forms when the torch moves too fast. The arc lags behind the cut, leaving a hard bead of uncut material or a rolled-over edge at the bottom. It is tenacious and usually has to be ground. Remedies: reduce speed in 125 mm/min steps, lower the standoff, increase amperage — but stay below about 95 % of the nozzle's rated current.

Top spatter is a light coating of re-solidified metal on the top surface, caused by a worn nozzle, excessive speed, or too high a standoff. It usually scrapes off, but it signals that the swirl angle has changed.

Between the two extremes is the dross-free window. It is a speed window, not a parameter set, and its width depends on the gas. Oxygen plasma on carbon steel has the widest window, because the exothermic reaction produces a fine, low-surface-tension spray that the jet ejects easily. Nitrogen and air plasma have narrower windows on steel. That is why a shop that switches from oxygen to air to save gas money often finds itself grinding dross it never had before.

The counterintuitive rule: dross usually means speed up, not slow down. Operators see dross, reduce feed, and convert high-speed dross into low-speed dross — a different defect with the same name. The cut face tells you which way to go. Slow cuts leave lag lines perpendicular to the plate; fast cuts leave S-shaped lag lines running along the bottom edge. In oxygen plasma, the correct speed gives a slight leading arc; in air plasma the arc should exit vertically; in nitrogen or argon-hydrogen, a slight trailing arc is correct.

Corners are a special case. The machine has to decelerate into a corner, which locally lowers the speed and produces low-speed dross. Modern controllers apply centripetal limiting and corner slowdown rules; if the corner dross is severe, the fix is usually in the CAM settings, not the console.

7. Holes, corners and minimum features

The hole rule is the most useful DFM rule in thermal cutting, and it is routinely ignored.

A conventional plasma torch cannot reliably cut a hole smaller than about 1.5 times the plate thickness. Below that ratio, the arc never reaches steady state, the hole becomes tapered and out-of-round, and the bottom diameter can be 1–2 mm smaller than the top. High-definition plasma pushes the limit to about 1.0:1; Hypertherm's True Hole process claims 0.75:1, and Kjellberg's contour-cut process has demonstrated 0.6:1 in 10 mm plate. Those are capability claims under specific conditions, not general drawing permissions.

The ratios mean:

Below the limit, the correct move is to cut a pilot hole and drill, or to drill from solid. A 12 mm hole in 15 mm plate is not a plasma hole; it is a drilling operation with a plasma-shaped roughing pass.

When a hole is within capability, the cutting parameters still need to change:

  1. Cut at 60 % of the contour feed. The slower speed reduces taper, at the cost of some low-speed dross on the bottom of the hole.
  2. Lead in at the centre of the hole, not on the circumference, so the torch has time to index from pierce height to cut height before reaching the radius.
  3. No lead-out. Overburn about 2.5 mm past the lead-in kerf, staying on the hole radius.
  4. Disable arc voltage control on holes below about 30 mm. The slower speed makes the THC think the torch is too high and drive it into the plate.
  5. Keep at least 1.5× thickness between the hole edge and the part edge, or the taper will be asymmetric.

The same logic applies to internal corners. A plasma kerf has a finite width, so a sharp internal corner becomes a radius of at least half the kerf, plus the lag. For a 2 mm kerf, expect a 1–1.5 mm internal radius. If the corner is functional — a bearing seat, a locating feature — draw a dogbone or a relief, not a sharp corner.

8. HAZ and what it does to the edge

The heat-affected zone is the part of the plate that did not melt but had its microstructure changed. Its width scales with heat input per millimetre of cut, which is why the four thermal processes separate so cleanly:

Process · HAZ width · Typical edge condition

Oxy-fuel · 4–10 mm · coarse grain, oxide, possible hardening

Plasma · 2–5 mm · recast layer, nitride/oxide, hardened rim

Laser · 0.2–2 mm · narrow recast, minimal distortion

Waterjet · 0 mm · no thermal effect

On 20 mm S355, an oxy-fuel cut can leave a 3.5–4.8 mm HAZ. If the carbon equivalent is high enough, that band hardens and becomes the initiation site for hydrogen cracking when the plate is welded without preheat. For S355 with CEV above 0.4, a 100–150 °C preheat is cheap insurance; for Hardox and Strenx, the plate maker's limits are stricter — 150 °C maximum for the 1100–1300 grades, 300 °C for the 700–960 grades — and the cutting speed must be kept up, not reduced, to avoid a brittle zone.

Plasma's HAZ is narrower but chemically dirtier. Air plasma nitrides the cut edge: one study of 14 mm S235JR measured a nitrided layer 80 µm deep at 300 mm/min, falling to 40 µm at 1,500 mm/min. That layer can cause porosity when the edge is welded. The remedies are to grind 0.1–0.2 mm off the edge before welding, to use F5 or nitrogen plasma gas instead of air, or to use a weld wire with deoxidising and denitriding additions.

Stainless and nickel alloys are their own problem. Plasma cutting vaporises chromium and can leave a chromium-depleted layer at the edge, reducing corrosion resistance; the cut face should be pickled and passivated if it will see service. On alloy 825, a 2025 study found that reducing the cutting speed increased heat input enough to form sigma phase and cracks in the HAZ — the opposite of the usual "slow down for quality" instinct. On 2205 duplex and 18-8L, the same study found HAZ widths of only 10–100 µm and no such sensitivity. The lesson is that "slow down" is not a universal quality lever; on some alloys it is a damage mechanism.

9. Gas selection and consumables

The gas decides the edge chemistry, the dross window, and the consumable life. The practical matrix:

Material · Plasma gas / shield · Result · Notes

Mild steel · Oxygen / air · Best cut quality, fastest, little dross · Standard for carbon steel; 800–1,500 starts

Mild steel · Air / air · Good, economical · Some nitriding and oxidation; weld prep may need grinding

Mild steel · Nitrogen / air · Fair, some dross · Excellent parts life; used when consumable cost dominates

Stainless < 12 mm · Nitrogen / air or CO₂ · Good quality, long life · CO₂ shield improves finish slightly

Stainless > 12 mm · Argon-hydrogen (H-35) / nitrogen · Excellent, near-polished · Hottest plasma; expensive; requires a gas-rated system

Aluminium < 6 mm · Nitrogen / nitrogen · Best quality · Air is acceptable but oxidises the edge

Aluminium > 6 mm · Nitrogen / water · Excellent finish · Water injection; not with H-35

Aluminium > 12 mm · H-35 or H-2 / nitrogen · Maximum capability · Never cut aluminium underwater with H-35 — hydrogen accumulation risk

Consumable life is a real cost line, but a smaller one than most shops think. Oxygen plasma sets run 800–1,500 starts on modern systems; nitrogen sets exceed 1,000 starts; air plasma sets on older machines can be as low as 600 starts. At 20 seconds of cutting per start, 1,000 starts is only 5.6 hours of arc-on time. A ₹1,100 consumable set therefore costs about ₹200 per arc hour — visible, but small next to the machine hour.

The safety side is not small. Plasma cutting stainless steel produces hexavalent chromium and nickel fume; air plasma produces ozone and nitrogen oxides; the arc is a high-intensity UV source; and the noise is 100–110 dB(A). A downdraft table or a water table, local extraction at the torch, and a welding helmet with a shade 8–10 filter are not optional in production. Oxy-fuel adds the fire risk of acetylene or propane, flashback arrestors, and the stored-oxygen discipline that goes with any high-pressure oxygen system.

10. The machine: THC, nesting, and the arc-on problem

A plasma table is a CNC positioning system, a torch height controller, and a gas delivery system. The height controller is the part that decides whether the other two produce good parts.

Arc voltage is proportional to standoff: raise the torch, voltage rises; lower it, voltage falls. The controller holds a voltage setpoint by moving the Z axis, which is why the cut charts list a voltage for every thickness and gas. The same charts assume a cut height of about 1.5 mm. If the plate is not flat, or the slats are burned through, the THC chases the voltage and the cut height wanders. A 1 mm height error is enough to change the bevel by a degree or two.

Piercing is the other machine-level decision. Plasma pierces quickly — about 1.5 s for 32 mm — but the pierce throws molten metal back at the nozzle. Pierce height is set at 2.5× cut height to give the spatter somewhere to go, and the pierce delay must be long enough for the arc to break through. On plate above about 40–45 mm, even a 300 A torch should edge-start: position the torch at the plate edge, start the arc there, and move into the cut. Trying to pierce 60 mm plate is how nozzles die.

Nesting affects cost more than most operators realise. Every pierce consumes a consumable start, and every unnecessary traverse is paid at the machine hour rate. On a hole-heavy part, the difference between a nest with 40 pierces and one with 120 is 80 starts, which is 8 % of a 1,000-start consumable set — and, more importantly, 80 opportunities for a pierce failure. Grouping parts by thickness and material, sharing a common cut line between adjacent parts, and sequencing holes to let the plate cool are all worth more than a 5 % speed tweak.

Then there is arc-on time. The cost-per-metre figures in the next section assume the torch is cutting 100 % of the hour. Real job shops run 30–60 % arc-on: the rest is loading, nesting, positioning, piercing, and waiting. At 50 % arc-on, every ₹/m figure doubles. This is why two shops with the same machine and the same cut chart can quote 40 % apart.

11. The rupee ledger: where the crossover actually is

The cost of a cut is dominated by the hour, not the metre. The formula is:

where c_{hr} is the fully loaded hourly cost in ₹/h and v is the cutting speed in m/min. The inputs used here are deliberately transparent, and a shop can substitute its own:

300 A HD plasma table

Oxy-fuel torch on the same gantry

At the chart speeds:

Thickness · Plasma speed · Plasma ₹/m · Oxy speed · Oxy ₹/m · Oxy, 2 torches ₹/m

25 mm · 1,950 mm/min · ₹35.9 · 430 mm/min · ₹106.6 · ₹55.8

50 mm · 560 mm/min · ₹125.0 · 300 mm/min · ₹152.7 · ₹79.9

60 mm · 385 mm/min · ₹181.8 · 300 mm/min · ₹152.7 · ₹79.9

100 mm · not capable · — · 230 mm/min · ₹199.3 · ₹104.3

The crossover is visible in the table. A single oxy-fuel torch is roughly three times more expensive per metre at 25 mm, draws level at about 55 mm, and owns everything above that. Two torches sharing the table and operator — the standard configuration on a profile-cutting gantry — bring the oxy cost down to ₹55.8/m at 25 mm and ₹79.9/m at 50 mm, which moves the economic crossover down to about 35–40 mm. With three torches it moves again, to roughly 30 mm. The crossover is not a property of the process; it is a property of the machine configuration and the hourly rate.

The breakdown is the more useful lesson. At 25 mm, plasma's machine hour is ₹22.2/m, electricity ₹5.37/m, consumables ₹7.63/m, and gas ₹0.68/m. The machine hour is 62 % of the ₹35.9/m total; power and consumables are another 36 %; gas is 2 %. At 60 mm, electricity rises to ₹27.2/m and consumables to ₹38.7/m, but the machine hour is still ₹112.6/m, or 62 %. The way to cut the cost of plasma cutting is to cut faster or keep the table busier, not to shop for cheaper nozzles. The same is true of oxy-fuel, where oxygen is under 3 % of the cost per metre; the 12.5 % speed penalty from a 0.5-point oxygen-purity deficit matters because it adds 12.5 % to the machine hour, not because the gas bill rises.

A worked example makes the purity tax concrete. A 25 mm cut at 430 mm/min covers 25.8 m/h, so at ₹2,750/h it costs ₹106.6/m. At 99.0 % oxygen, speed falls by 12.5 % to 376 mm/min, or 22.6 m/h, and the cost rises to ₹121.8/m — ₹15.2/m, or 14 %, on every metre of plate. Over a 2,000 m annual program, that is ₹30,400 of pure loss from a gas supply nobody tested.

12. What to put on the drawing

Thermal cutting is a process with real capability limits, and a drawing that states them gets a realistic quote. The items worth adding:

  1. Material grade and thickness, including the standard — "S355J2 to IS 2062 E350", not "MS plate".
  2. Cut quality standard: "ISO 9013 Range 3" or "Range 4" on the cut face, rather than an unqualified ±0.5 mm.
  3. Hole schedule: holes at or above 1.5× thickness may be plasma-cut; smaller holes are drilled. State which is which, or let the shop decide and price it.
  4. Critical dimensions and the datum they are measured from. Plasma is an IT12–IT13 process; do not put an IT9 tolerance on an as-cut edge without a finishing note.
  5. Weld prep: whether the cut edge will be welded as-cut, ground, or machined. This decides whether the nitride/oxide layer is acceptable.
  6. HAZ-sensitive areas: hardened or high-strength steels, and any restriction on preheat or cutting speed.
  7. Flatness and distortion limits, especially for thin sheet and long slender parts.
  8. Edge condition on visible or sealing faces — dross-free, ground, or machined.
  9. Quantity and nesting preference, because the difference between 6 parts and 600 changes the process choice.
  10. The 3D model or DXF, not a PDF alone. Kerf compensation, lead-ins and hole sequencing are CAM decisions that need geometry, not a picture.

On FabFlow, jobs that carry the material standard, the cut-quality range, and the hole schedule attract quotes that are both lower and more accurate, because the manufacturer is not pricing three different assumptions.

13. The short version

Plasma and oxy-fuel are mature processes, and their physics has not changed in decades. What changes is the price of getting the details wrong: a 1.2:1 hole in 15 mm plate is not a small error, it is a scrapped bracket with 1.2 mm of taper, a hard dross bead, and a drill operation that was never costed. The arithmetic in this guide is the difference between quoting that part correctly and discovering the error at the welding table.

More FabFlow blog posts