What oxygen pressure and cutting speed should you set for CNC flame strip cutting?

What oxygen pressure, nozzle size and cutting speed should I use for 12-40 mm strips?

Start from the thickness band, then trim with a trial cut. For 10-25 mm mild steel use a 1.2-1.6 mm cutting oxygen orifice (nozzle #3), cutting oxygen 0.4-0.6 MPa at the torch inlet, stand-off 3-4 mm and 350-500 mm/min; kerf lands around 2.4-2.7 mm. For 25-40 mm go to a 1.6-2.0 mm orifice (#4), 0.5-0.7 MPa, 4 mm stand-off and 320-400 mm/min, kerf 2.8-3.2 mm. For 40-60 mm use 2.0-2.2 mm (#4-5), 0.6-0.8 MPa, 5 mm stand-off and 250-320 mm/min. These figures assume clean S235/S355 plate, 99.5% oxygen and a neutral preheat flame, and they are a starting point only: machine builders' charts differ by 10-20% because torch design and hose pressure drop differ. The reliable check is the drag line - at the correct speed the drag lines on the cut face are almost vertical, 0-3 degrees. Drag leaning back means the machine is running too fast or oxygen pressure is low; a nearly horizontal drag with heavy dross means too slow. On the YOMI YM-CS the F2300B controller plus StartCAM stores one parameter set per thickness and grade, so the correct values are recalled rather than re-entered for every job.

How do I set the preheat flame and stand-off so the top edge does not round over?

Set the preheat flame neutral to very slightly reducing, and keep the inner cone 3-5 mm above the plate surface - never let the bright cone touch the steel. Flame core touching the plate causes the classic rounded, melted top corner (and can carburise the cut face). Stand-off must follow thickness: about 3-4 mm on 12-25 mm plate, 4-5 mm on 30-60 mm, 5-7 mm at 80-100 mm. Too high spreads the jet and rounds the upper edge while widening the kerf; too low produces molten droplets stuck along the top edge and slag beads. If the top edge shows a melted overhang, the usual causes in order are: preheat oxygen too high, nozzle number too large for the thickness, travel speed too slow, then stand-off wrong. If the cut face is concave or wavy in the middle, the cause is normally cutting oxygen pressure too high or a partially blocked nozzle. Since the YM-CS runs multiple torches across a 3150 mm width, every torch rides its own 400 W Panasonic servo Z axis with torch height control, which keeps stand-off identical on all strips - an inconsistent flame on one torch shows up immediately as a different edge on one strip.

Should I pierce the plate or start each cut from the plate edge?

On strip work, start from the edge wherever the geometry allows. Position the torch 2-5 mm inside the plate edge, preheat until the spot is bright orange, then open the cutting oxygen and move off - no pierce, no blowback into the nozzle, less dross, and noticeably longer nozzle life. Piercing is only needed for internal cutouts and for transverse cuts that begin inside the plate. Piercing preheat time scales with thickness: roughly 6-8 s at 12-20 mm, 8-10 s at 25-40 mm, 10-15 s at 50-60 mm and 25-40 s at 80-100 mm. Practical rules: pierce 8-12 mm outside the finished contour and lead in, lift the torch 3-5 mm above normal stand-off during the pierce and drop back before the cut starts, and on plate over 60 mm use an oscillating pierce so the oxygen jet is not blowing into a blind hole. Most nozzle failures are pierce blowback, not cutting hours, so a shop that converts internal pierces to edge starts typically doubles nozzle life. Track pierce count per shift - it is a better consumable indicator than cutting hours.

Do I need a liquid oxygen tank, a cylinder manifold or a PSA generator for a 9-torch machine?

Size by peak flow, not by average. A single torch on 25-40 mm plate draws roughly 3-6 m3/h of cutting oxygen plus 0.5-1.5 m3/h for preheat, so nine torches firing together pull 30-55 m3/h, and that peak arrives all at once when the plate is lit. A standard 40 L cylinder holds about 7 m3 of usable gas, which means 5-8 cylinders per hour at full duty - a cylinder pallet is only practical for low duty or a 1+1 single-torch configuration. Three supply routes: (1) liquid oxygen tank with an ambient vaporiser, 1-3 m3 (about 800-2400 m3 of gas) giving one to two weeks per fill at 99.5% purity - the normal answer for multi-torch strip production; (2) a cylinder manifold with automatic changeover for one or two torches or intermittent work; (3) a PSA or membrane generator at 93-95% purity, which is cheaper per m3 but costs cut quality (see the purity question). Whatever the source, pipe sizing matters as much as the tank: use a 3/4 inch header minimum, 1 inch above eight torches, and a loop or ring main so pressure does not droop at the far end of the gantry.

How much does oxygen purity actually affect flame cutting quality?

A lot, and it is the single most under-checked variable in strip shops. Take 99.5% as the benchmark the cutting charts assume. At 99.0% you lose roughly 5-8% of cut speed and start seeing slightly heavier dross. At 97% the loss is typically 15-25% of speed, the kerf widens 0.3-0.5 mm, and the dross becomes hard and adherent, needing chipping or grinding before the strip goes to assembly. At 93-95% - what a PSA generator delivers - thin plate can still be cut, but 60 mm and above becomes unreliable and the extra oxygen you burn trying to compensate erases much of the cost saving. Two other points: measure purity at the torch inlet, not at the tank or the generator outlet, because leaks and contaminated hose assemblies pull purity down along the way; and check moisture and dew point too, since wet oxygen causes dross, nozzle carbon build-up and erratic preheat. Before blaming the machine for a quality change, sample the gas - it explains a whole class of 'the machine cuts differently than it did last month' complaints.

Why do strips at one end of the gantry cut differently from those at the other end?

Almost always a gas distribution or thermal issue, not machine accuracy. Four causes, in order of frequency: (1) asymmetric gas manifold - if the header feeds from one end, the last torch sees both lower pressure and slower response when all torches open at once, so fit a ring main or per-torch regulators with gauges; (2) mismatched hose bore and length - a 6 mm hose on one torch and 8 mm on another gives different flow under peak demand; (3) plate temperature - by the third pass the middle of the plate is several hundred degrees warmer from the previous cuts, so the last strips need about 3-5% higher speed or they overheat the top edge and produce wider kerf; (4) condition differences - mill scale, rust or paint on one end of the plate changes preheat time and edge quality. Fixing order: standardise hose bore and length, convert the header to a loop, fit one gauge per torch so a droop is visible, stagger torch ignition slightly if supply is marginal, and cut the outside strips before the plate heat builds.

How do I calibrate parameters for a new plate grade or thickness and keep them?

Run a controlled trial cut on an offcut of the same grade and thickness, 300 mm long, and measure four things: kerf width at top and bottom, drag angle, dross on the bottom edge, and top-edge roundness. Start at 90-95% of the book speed and adjust one variable at a time - speed first in 10% steps, then cutting oxygen pressure in 0.05 MPa steps, then stand-off in 1 mm steps. Changing two variables together makes the result unreadable, which is why so many shops never converge. Then write the accepted set into the F2300B / StartCAM parameter library keyed by thickness plus grade plus nozzle number, and keep a laminated chart at the machine listing thickness, nozzle, oxygen pressure, speed, stand-off and preheat time. Re-verify whenever something in the chain changes: new nozzle brand, new oxygen supplier or a purity drop, a different plate mill, or a seasonal humidity swing - wet oxygen in summer is a classic cause of 'the same recipe stopped working'. Because the YM-CS stores recipes in the CNC, calibration effort is paid once per material rather than once per shift.

What changes when I move from 20 mm to 80-100 mm plate?

Everything scales up, and some things change qualitatively. Nozzle goes to 2.2-3.0 mm (#5-6), cutting oxygen to 0.6-1.0 MPa, stand-off to 5-7 mm, and speed drops to about 220-270 mm/min at 80 mm and 180-220 mm/min at 100 mm; preheat time extends to 25-40 s because you are heating a much larger mass before the oxygen can start the reaction. Above roughly 40 mm, divergent (Laval) tips beat cylindrical tips: the jet stays coherent through the full depth instead of spreading, giving a squarer edge and lower oxygen consumption per metre. Fuel gas becomes the constraint - acetylene is limited to about 1 bar gauge and cylinder draw rates cannot feed a multi-torch gantry at these thicknesses, so propane or natural gas with a two-piece nozzle set is the practical choice. Expect 3.4-3.5 mm kerf and 1-2 mm of edge draft from the natural top-wide/bottom-narrow tendency, so plan machining or bevel allowance if the strip edge is a fit-up face. Handling also changes: 100 mm plate at 3.5 m width is over 8 t, so lifting gear, roller support and strip tipping control matter as much as the cutting parameters, and long thick cuts put enough heat in the strip to justify skip cutting with bridges left in place until the plate cools.

Technical Specifications of CNC Flame Strips Cutting Machine

Product Gallery

Applications

The CNC Flame Strips Cutting Machine is widely applied in steel structure prefabricated buildings, bridge and infrastructure construction, shipbuilding and marine engineering, heavy machinery manufacturing, power plant and industrial facilities, and offshore wind power structures.

Learn more about the CNC Flame Strips Cutting Machine:
https://www.steelstructurer.com/pid18376093/CNC-Flame-Strips-Cutting-Machine.htm

For more information about CNC Flame Strips Cutting Machine and other steel structure intelligent equipment, contact YOMI CNC Cutting & Welding Machinery. Our team of experts is ready to help you find the best solution for your manufacturing needs.