Bending is the operation every duct shop underestimates. Lockformers, flange formers and shears get the capital budget, while brackets, hanger angles, stand frames, duct corners and specialty fittings keep coming off a tired hand brake — one bend at a time, by feel. A CNC press brake closes that gap: it stores the bend program once and repeats the angle and flange length to within ±0.2° for as long as the job runs. This guide explains how the machine works, how to size tonnage against your real material mix, how to pick V-die openings, and what to check before you sign the acceptance report.
A press brake bends sheet between two matched tools: a punch clamped to the moving ram and a V-die sitting in the bed. The CNC controller drives the hydraulic ram to a programmed depth and holds the backgauge at a programmed position, so the bend line is placed by the machine instead of by chalk marks. Because ram depth and gauge position are stored in the part program, the hundredth bracket is identical to the first — that repeatability is the entire business case for the CNC class.
Modern hydraulic brakes control each cylinder independently as the Y1 and Y2 axes. A closed-loop encoder on each side reports ram position, and the controller keeps the beam parallel to the bed through the whole stroke. Older torsion-bar machines link the two cylinders mechanically; they are cheaper and still accurate on short beds, but electro-hydraulic Y1/Y2 sync holds tighter parallelism on wide beds and corrects small mismatches automatically. When you compare quotations, ask which sync method the quoted model uses — it is one of the few internal differences that shows up in real-world angle consistency.
Almost all HVAC work is air bent: the punch presses the sheet into the V without pinching it to the die walls. The angle is set by ram depth, so one die set covers a range of angles, and springback is compensated by bending a fraction deeper. Bottoming and coining push the sheet into full contact with the die for near-zero springback, at the cost of exact tonnage control and matched tools per angle. For 0.5–2 mm galvanized and stainless fittings, air bending with programmed depth compensation is faster and far more forgiving of material batch variation.
Tonnage decides what you can bend, and bed length decides what fits. The manufacturer's press brake range runs from compact 400 kN (40 T) machines with 2,200 mm tables up to 6,000 kN frames with 8 m beds, but duct and HVAC fabrication lives in the 40–100 T window — the class most export quotations configure to order.
For air bending mild steel, the working rule is a V-die opening of roughly 8× the material thickness. Bend a 2 mm blank in a 16 mm V and the tonnage demand stays moderate; squeeze the same blank into an 8 mm V and required force rises steeply, punishing the tooling and the frame. The rule exists to keep you in the low-force zone: when a sales table shows tonnage per meter, it assumes the 8× opening. If your parts force tighter dies, re-compute instead of trusting the table.
Published tonnage charts assume mild steel around 450 N/mm² tensile. Stainless steel needs roughly 50% more force at the same thickness, and aluminium needs about half. A shop bending 1.5 mm stainless angle iron should therefore quote the stainless case first and let mild steel fall out as free capacity. Remember also that capacity is per bend length: a 40 T machine bending a 1 m bracket is a different animal from forcing the same 40 T across a full 2.5 m bend.
| Model | Nominal force | Worktable | Throat depth | Ram stroke | Daylight | Main motor | Weight |
|---|---|---|---|---|---|---|---|
| 40/2200 | 400 kN (40 T) | 2,200 mm | 200 mm | 100 mm | 330 mm | 4 kW | 2,900 kg |
| 63/2500 | 630 kN (63 T) | 2,500 mm | 250 mm | 120 mm | 360 mm | 5.5 kW | 4,500 kg |
| 80/3200 | 800 kN (80 T) | 3,200 mm | 320 mm | 120 mm | 380 mm | 7.5 kW | 6,600 kg |
| 100/3200 | 1,000 kN (100 T) | 3,200 mm | 320 mm | 120 mm | 380 mm | 7.5 kW | 7,200 kg |
| 125/3200 | 1,250 kN (125 T) | 3,200 mm | 320 mm | 130 mm | 390 mm | 11 kW | 9,200 kg |
| 160/3200 | 1,600 kN (160 T) | 3,200 mm | 320 mm | 200 mm | 450 mm | 11 kW | 11,700 kg |
The model code reads tonnage over bed length: a 63/2500 delivers 630 kN across a 2,500 mm worktable. Stroke figures matter more than buyers expect — a 100 mm stroke class handles shallow brackets and flanges, while 160–200 mm stroke models give room for deep boxes and tall stands. Column distance limits the widest die assembly you can swing between the frames, and throat depth determines how far into the sheet an edge flange can sit.
Keep a die set per material family: 6–8× thickness for galvanized mild steel, and lean to the larger end as thickness or hardness grows. The inside radius produced by air bending is roughly 1/6 to 1/8 of the V opening on mild steel, which is why the same 16 mm die bends 2 mm steel with an acceptable radius but leaves marks on soft aluminium — for aluminium, widen the V or use polyurethane pads. Minimum flange length is governed by the die shoulder: the flange must span at least the V opening plus clearance, typically V × 1.2, or the sheet will fall into the groove and the bend will drift.
Under load both ram and bed deflect, and the middle of a long bend comes out open. NC and CNC machines solve this in one of two ways: a motorized crowning table that pre-cambers the die upward in the center, or a hydraulic deflection compensation system that adjusts with each stroke. For duct-shop bend lengths — under 3 m and under 100 T — a manual or NC-controlled crowning system is usually sufficient; full hydraulic crowning earns its cost on heavy tonnage and long beds. Whatever the mechanism, the acceptance test is the same: bend a full-width test piece and measure the angle at both ends and the center. A spread of more than ±0.5° across a 3 m bend is a crowning problem, not an operator problem.
The backgauge is what turns a press into a production machine. CNC gauges run on ball screws with encoder feedback, typically positioning flange lengths in 0.1 mm steps with finger stops that can be individually raised or lowered along the X axis. Better controllers add more axes — R (gauge height), Z1/Z2 (independent finger position), and delta-style extensions for hemming and offset work. For a duct shop, X-axis accuracy plus two or three fingers covers brackets, hanger angles, cleat blanks and corner fittings; paying for five or six axes only pays if you run complex profiles daily.
The daily workflow is short: recall the part program, load the sheet against the fingers, foot-pedal stroke. Angle control is closed-loop on the Y1/Y2 axes — the controller measures ram depth and, on higher-spec NC systems, corrects for measured springback per material batch. One practical buying question: can the operator store a new program in under a minute without a programmer? If setup takes an hour, small fitting batches will quietly migrate back to the hand brake and your payback calculation dies.
The two machines overlap less than buyers assume. A box and pan brake — the 2,500 mm manual finger brake covered in our separate guide — clamps the sheet and folds boxes, pans and light fittings with zero tonnage math, but it is manual and one bend at a time. A press brake does not make long crisp box folds as efficiently, but it handles anything needing force: 2 mm and heavier brackets, stainless hangers, angle frames, thick cleat stock. Shops running an auto duct line with a lock former and a TDF flange former typically add a 40–100 T CNC press brake as the fittings cell and keep the pan brake for lightweight pans. If your cutting is still manual, pair the brake with a hydraulic guillotine shear so blank accuracy matches bend accuracy.
Run this list before the container leaves the factory floor — every item is cheaper to fix before shipment than after installation:
| Symptom | Most likely cause | What to do |
|---|---|---|
| Angle drifts through the batch | Material batch springback change; no depth correction | Re-teach ram depth on the new batch; on NC systems, log springback per material code |
| Middle of long bends opens | Frame deflection; crowning not set | Program crowning for the bend length; verify with the full-width test piece |
| Flange length varies part to part | Backgauge ballscrew backlash or loose finger stops | Check gauge repeatability with a caliper; tighten finger clamps, report ballscrew play under warranty |
| Hydraulic oil overheats | Cooler fouled or pump holding relief pressure | Clean the cooler; confirm the controller idles the pump instead of holding relief all shift |
| Tooling marks on stainless | Die shoulders worn or contaminated, V too tight | Widen the V, clean and polish shoulders; keep separate die sets per material family |
| Ram stalls mid-stroke on thick work | Job exceeds tonnage at the chosen V opening | Recompute tonnage for the actual V; do not force the job — a stretched frame never holds angle again |
For the fittings mix a duct shop actually bends — 0.5–2 mm galvanized brackets and angles, occasional 1.5 mm stainless — 40 to 100 T is the working window. Size on your heaviest stainless job at the correct V opening, not on the thickest mild steel you have ever been handed once.
Yes for nearly all fittings: hanger angles, stands, corner reinforcements and cleat stock rarely exceed 2.4 m. Choose a 3.2 m bed only if you bend long architectural flashings or full-length stiffeners.
Yes — allow roughly 50% more tonnage than the mild steel chart and use a wider V opening to limit tool pressure. Confirm the frame and hydraulic rating are quoted for the stainless case.
An NC system with programmable Y1/Y2 depth, X-axis backgauge and program storage covers duct-shop work. Higher-end colour-screen controllers add axis expansion and sheet-count logging — nice, not decisive. What matters is that your operators can build a program in minutes.
Export-class CNC hydraulic brakes in the 40–100 T window are quoted around $18,500 configured to order, moving up with tonnage, bed length and controller specification. Get the tonnage and bed fixed first; controller features are the cheapest part of the quote to trim later.