Two cylinders, one commanded Y value. As long as the workpiece is centred, the two rams travel together and the error stays hidden. Put a heavy plate at one end of a 4 m bed, or bend a narrow bracket 1.8 m from the centre line, and the two sides of the ram no longer see the same load. What the machine does next decides whether the flange positions are within tolerance.

1. The practical problem

The symptom appears as a wedge across a part that should be parallel: one end of the bend angle is higher than the other, or a box comes out with flanges of different height. It disappears when the same part is run centred, and it worsens as the load moves further from the centre line or as plate thickness increases. Operators often compensate with shims or by adjusting the crowning, which masks the fault and makes the next job worse. On machines without independent position feedback, the same error also drifts with oil temperature, so the part can be good in the morning and out of tolerance in the afternoon.

2. Failure mechanism

Each cylinder is driven by its own valve. When the load is asymmetric, the more heavily loaded cylinder needs higher pressure to move at the same rate. Oil at higher pressure is slightly less compressible, so its flow-versus-command relationship changes and it lags behind the lightly loaded side. Valve dead band, hysteresis and spool overlap add a command error on top. The result is a ram that does not descend as a straight horizontal line: it tilts, settles, and in some cases corrects itself as pressures equalise, which is why the error can be intermittent. A machine with a mechanical torsion bar tying the two sides limits the tilt structurally but cannot eliminate it, because the bar itself twists under load.

3. Key parameters

The relevant parameters are the control architecture, the feedback resolution, the pressure difference between the two cylinder circuits, the load eccentricity, and the beam and ram stiffness. Position feedback is decisive: a closed loop on each ram measures what actually happened and corrects it, while an open-loop command assumes both sides obeyed. Pressure difference is the most useful on-machine indicator, because it is proportional to load asymmetry and is available in most controllers or from two gauges.

4. Diagnosis and calculation method

• Read the two cylinder pressures with the load centred. A significant steady difference with a centred load indicates a valve, seal or preload problem, not an eccentricity problem.
• Run a three-position test on a 1000 mm strip: centred, 300 mm left, 300 mm right, at the same programmed Y.
• Record the angle at both ends of the strip in each position; the change between positions is the synchronisation error expressed in angle.
• Repeat the test after two hours of production. Growth over time indicates a thermal contribution and should be separated from the mechanical error before adjustment.
• Where scales are fitted, compare Y1 and Y2 readouts during the stroke, not only at the bottom; the difference during descent is the real indicator.
• Convert angle difference to positional difference using the die geometry, or simply record both, since the machine's own thickness and die data give the ratio.

5. Prevention

Specify dual-scale feedback when work involves long or narrow parts bent off centre; it is the only architecture that measures the error it is correcting. Keep the hydraulic oil within its working temperature band, since viscosity change amplifies the difference between the two circuits. For heavy structural work, a stiff ram and a wide distance between frames reduce the eccentric moment itself. Our 160T/4000 synchro servo press brake uses independent servo-controlled valves with a linear scale on each side, and heavier structural bending is handled by our 400 tonne press brake for steel beam work, where the monoblock frame is designed for asymmetric loads.

6. FAQ

Does this affect the bend angle or the flange length?

Both, but through different routes. Ram tilt changes the angle progressively along the bend, while the same tilt shifts the position where the sheet meets the die, which moves the flange. On a four-bend box the flange effect accumulates and is usually noticed first.

Will a crowning adjustment fix an off-centre wedge?

No. Crowning acts symmetrically about the centre line. Applying it to an asymmetric error raises the middle of the bed and leaves the wedge in place, while disturbing the parts that were previously correct.

Can I test this without instrumentation?

Partially. A three-position strip test with a digital protractor shows the effect, but the cause can only be separated with pressure readings or Y1/Y2 position data. Without those, the result distinguishes "present" from "absent" rather than identifying the mechanism.