The machine passed its factory test, was dismantled, shipped, reinstalled and now bends differently from the recorded result. Nothing is visibly damaged. Between the two measurements four conditions changed: the foundation, the residual stress state of the frame, the alignment of major components after reassembly, and the calibration of the position feedback.

1. The practical problem

The complaint arrives as a general loss of accuracy rather than a specific fault: angles that need correction, flanges that no longer line up, a machine that needs a different crowning setting for the same job. Sometimes the error is a constant offset that is easy to correct; sometimes it is a wedge that cannot be corrected at all. Distinguishing the two early prevents a long argument about whether the machine or the installation is responsible.

2. Failure mechanism

A press brake frame is stiff in bending but not rigid in torsion. If the floor is not flat and level within the installation tolerance, bolting the frame down imposes a twist that was not present at the factory, and the ram-to-bed relationship changes along the length. Welded frames also carry residual stress; after transport, re-levelling and the first weeks of loading, that stress redistributes and the frame settles by a small but measurable amount. Reassembly introduces its own errors: scale brackets stressed in transit, backgauge rails not square, hydraulic pipes refitted with different preload. Each of these has a different geometric signature. The effect of a twist is small in absolute terms and large in angle terms, because a few hundredths of a millimetre across the bed changes the load distribution and therefore the penetration that decides the angle.

3. Key parameters

The parameters to record are floor levelness in both directions, the ram-to-bed gap at both ends, the parallelism of the ram over its stroke, the backgauge squareness to the bed, the scale calibration offset at both ends, and oil temperature at the time of measurement. Recording them at installation and again at thirty days provides both the correction data and the evidence base. Two of them are worth emphasising because they are cheap and decisive: the precision level survey, which shows whether the frame is being forced into a twist, and the ram-to-bed gap at both ends, which shows whether the machine is parallel before any hydraulic correction is applied. If both are correct and the angles are still wrong, the fault is in the compensation setting or in the tooling, and the diagnosis moves on rather than repeating the same level check.

4. Diagnosis and calculation method

• Level the machine in both directions before and after tightening the anchor bolts; tightening alone can move a foot by a measurable amount.
• Check free movement of the ram by hand over the full stroke for binding, which would indicate frame distortion.
• Bring the ram onto gauge blocks at both ends and compare indicated position with the physical gap to establish the scale offset.
• Set the backgauge zero against a machined reference and verify on a part of known flange length.
• Repeat the factory angle test at two or three tonnage levels, so the result can be compared like for like.
• Repeat the whole set after thirty days of production; a single settling step is normal, continued movement is not.

5. Prevention

Prepare the foundation before the machine arrives, level it with a precision level rather than a builder's spirit level, and keep the factory test record as the reference document for the acceptance test on site. Specify lifting points and never sling under the bed or on the guideways. Where the machine is built with a stress-relieved monoblock frame machined in a single set-up, settling is smaller and more predictable; our 160T/4000 synchro servo press brake and the 160 tonne hydraulic press brake range are supplied with installation and re-test instructions for exactly this reason.

6. FAQ

Is it normal for accuracy to change at all?

A small, one-time change in the first weeks is normal and expected after transport and re-levelling. Continued change indicates a foundation, loading or assembly problem that should be investigated.

Do I need to recalibrate the scales after installation?

Verify them, yes. Compare the indicated position with gauge blocks at both ends and correct the offset if required, always with the machine levelled and at working temperature.

Can I skip the thirty-day re-test?

It is the cheapest test in the machine's life. Without it, a settling trend is discovered only when parts start failing, at which point the original baseline no longer explains anything.