Hydraulic or servo busbar bending

Ask a machine supplier whether you should buy hydraulic or servo and you will get back a description of whichever one they build. That is not useful to anyone.

We build both. The EMAC-BB-H12 is a hybrid hydraulic head and the EMAC-BB-S12 is a pure servo one. They are rated identically: 400 kN, ±0.1° bending accuracy, 200 × 16 mm capacity, electronic springback compensation, 6.25 kW installed. Same envelope, same specification sheet, different drivetrain. So the choice between them is a choice about how you run your shop rather than about which machine is better, and we have no commercial reason to pretend otherwise.

Three architectures, not two

The comparison is usually posed as two options. It is three, and the middle one is where most of the current confusion lives.

On a conventional hydraulic head, a fixed-displacement pump on an induction motor runs continuously at full speed. Flow the machine does not need goes over a relief valve and becomes heat in the oil. Ram motion is controlled by directional and proportional valves, and the unit carries an accumulator, a sizeable reservoir and usually a cooler.

On a hybrid, or servo-pump, head, a servo motor drives the pump and pump speed is commanded to match demand. Ram position is controlled by turning the pump rather than by throttling flow across a valve, so the relief path stops being the primary control element. Because the losses are much smaller, the reservoir shrinks and the cooler often disappears. Hydraulically it is still a hydraulic machine: cylinder, oil, seals, filter.

On a pure servo head, a servo motor drives a ball screw, roller screw or rack, and there is no working fluid at all. The only oil in the machine is gearbox and guideway lubricant.

When people say "servo bending machine" they sometimes mean the second and sometimes the third, and the two behave very differently on force and on maintenance. Ask which one a quotation is describing.

Force at full stroke

This is the structural advantage of hydraulics and it does not get discussed enough.

A hydraulic cylinder produces force equal to pressure times piston area, and that relationship does not care where the ram is. Full rated force is available at the start of the stroke, at the end, and everywhere between. It is also available statically: you can hold at pressure, indefinitely, with no thermal penalty in the drive.

A screw-driven servo axis produces force from motor torque through the screw lead, and while that is also nominally constant with position, the practical limits are different. Holding high force at zero speed puts current through motor windings with no rotation to help cool them, so continuous-duty force is lower than peak force. Screw and nut life is governed by a load-times-distance integral, so working near maximum force continuously consumes the drivetrain faster than working at half force. And the whole train, screw, nut, bearings and frame, has to be sized for the peak, which is why a servo machine at a given force rating is generally the more expensive of the two to build.

For thin flatwise bending at 5 to 10 mm this rarely binds. For thick edgewise work at the top of a machine's capacity, where the bar is at full section and the tooling is at full engagement for a long dwell, a hydraulic cylinder is doing what it was designed to do and a screw is working hard.

Energy: idle is where the difference lives

Under load, the two architectures are not very different. A hydraulic system loses something to pump and valve efficiency and to fluid friction, a servo screw loses something to motor efficiency and screw friction, and neither loss dominates the picture.

Idle is where they diverge, and busbar bending is a high-idle process. The operator repositions the bar between bends, checks a dimension, walks to the rack for the next length. On a conventional hydraulic machine the pump turns at full speed through all of that, dumping flow over the relief valve and heating the oil. On a servo pump the motor slows or stops. On a pure servo axis the motor draws holding current only.

Quantifying it is harder than the marketing suggests. Suppliers of variable-speed pump drives commonly claim 30 to 60% energy reduction against fixed-speed units, and some claim up to 80%, always with the caveat "depending on duty cycle" doing a great deal of work. Those are vendor figures and they are not independently verified here.

The closest thing to independently measured data comes from an adjacent machine class. A US utility programme monitored a hydraulic and an all-electric injection moulding machine side by side for a week on comparable parts, and reported 0.278 kWh per pound of plastic for the hydraulic machine against 0.073 kWh per pound for the all-electric one. That is a different process with a different duty cycle and it should not be transferred to busbar bending as a number. It is here only because the direction and the rough scale hold up under measurement rather than under marketing.

For your own case, do the arithmetic rather than trusting a percentage. Installed motor power on both our heads is 6.25 kW. Take your shift length, subtract the time the ram is actually moving under load, and price the remainder at your industrial tariff. In a one-shift shop with modest utilisation, the annual difference is usually small enough that it will not decide the purchase. In a three-shift busway plant, it is large enough to appear in a payback calculation, and it comes with a second saving that people forget: energy that does not go into the oil is energy you do not then pay to remove with a cooler, and heat you do not add to the shop in summer.

Accuracy and repeatability

Both heads are specified at ±0.1° with electronic springback compensation, and that is not a marketing coincidence: on a modern machine, positional accuracy is set by the measurement system and the control loop, not by what pushes the ram.

Where the architectures differ is in what the loop has to fight.

A servo screw has a stiff, near-linear relationship between motor position and ram position, with backlash and screw wind-up as the main errors, both of which are stable and can be compensated. A hydraulic ram has an oil column between the pump and the load, and oil is compressible. Its stiffness changes with pressure, with temperature and with entrained air. A closed loop on ram position handles this perfectly well, but the loop is working harder and the dynamic response is softer.

The practical difference shows up in short, fast moves rather than in final accuracy. A servo axis settles faster from a small commanded increment, which matters when a programme is making many small angle corrections. On a single 90° bend to a stop, you will not be able to tell them apart on a protractor.

Springback compensation is a control function and lives in the CNC on both. It works from a stored correction per material, temper and tool combination, which is a materials problem rather than an actuator problem, and is dealt with separately in springback compensation in busbar bending.

Noise

A conventional hydraulic power unit is the noisiest object in most busbar shops. The noise is dominated by the pump, and it is continuous, which is more fatiguing than intermittent noise at the same level.

Both variable-speed hydraulics and pure servo cut this substantially, and for the same reason: the pump either slows down or does not exist. A pure servo machine at idle is effectively silent, and its noise under load is mechanical, from the tooling contacting the bar, which is the same on any machine.

We are not going to quote a decibel figure, because the number depends on the enclosure, the mounting, the reservoir, the room and where you stand, and any single figure would be meaningless. If noise is a constraint for you, ask for a measurement at a stated distance under a stated duty cycle, from any supplier, and treat a refusal as an answer.

Oil, and what it actually costs

Hydraulics bring a maintenance stream that has nothing to do with bending. Filters on a schedule. Oil analysis, or at least oil changes on a schedule. Hose and seal replacement, which is not scheduled and therefore always inconvenient. A reservoir that has to be topped up, kept clean and eventually disposed of as controlled waste.

There is also a housekeeping issue specific to this industry. Busbar goes on to be insulated, plated, powder coated or encapsulated, and hydraulic oil mist on a copper surface is a coating adhesion problem. Shops that run hydraulic bending next to a coating line already know this and have a degreasing step. Shops that assume it will be fine find out at the first batch of blistered bars.

The hybrid architecture reduces this without removing it. Less heat in the oil means longer fluid life and a smaller reservoir to change, but the seals, hoses and filter are still there.

A pure servo machine trades this for a different, smaller list: screw and guideway lubrication, backlash checked periodically, and a drivetrain whose eventual rebuild is a bigger single event than any hydraulic service but happens far less often.

Cold start

An underrated difference in unheated shops.

Hydraulic oil viscosity is strongly temperature dependent. Cold oil moves slowly, and a machine started at 5 °C on a winter morning behaves differently from the same machine at its steady-state temperature: cycle times are longer, and until the oil column and the frame reach equilibrium, the first parts of the shift can drift on angle. The usual answer is a warm-up cycle, which costs shift time and energy, and it is one of the reasons a hydraulic shop tends to leave machines running through breaks.

A servo axis has no equivalent. Grease viscosity changes slightly and the frame still moves thermally like any machine tool, but there is no working fluid whose bulk modulus and viscosity are in the control loop. First part off is closer to correct.

The hybrid sits in between and closer to conventional hydraulic on this point, because it still has oil, though a smaller volume of it that spends less of its life hot.

Cost

We are not going to publish prices, and any article that gives you a price ratio is guessing about your specification, your country and your year. What we can give you is the shape of the cost.

Purchase price for a given force rating usually runs pure servo above hybrid hydraulic above conventional hydraulic, because a screw train sized for the full force is expensive metal.

Energy cost is the reverse and it scales with utilisation and with your tariff. At low utilisation it is noise in the accounts; at three-shift utilisation and European or Japanese electricity prices it is not.

Maintenance cost is a scheduled, small, recurring amount on hydraulics against an unscheduled, larger, rarer amount on servo. Which of those you prefer is partly a financial question and partly a question of whether you have a maintenance department.

Downtime is where the money usually is, and it does not favour either architecture consistently. A hydraulic machine has more things that fail gradually and can be caught. A servo machine has fewer failure modes but a longer parts lead time when a drive or a screw does go.

The decision rule

Buy the pure servo head if any of these describes you: you run high utilisation across two or three shifts; your product mix is dominated by repetitive small and medium bends where cycle time and settling time matter; you are in a clean or coating-adjacent environment where oil is a nuisance; noise is a constraint; or you are at the thin end of the capacity range and will never work the machine near its force limit.

Buy the hybrid hydraulic head if: you regularly work at the top of the capacity, particularly thick edgewise bends and long dwells; your duty cycle is bursty, with periods of heavy work separated by idle time, which is exactly what the servo pump is designed for; you have a maintenance team already competent with hydraulics; or you want hydraulic force behaviour without a conventional power unit's standing energy bill.

Buy conventional hydraulic if the machine will be lightly used, the capital budget is the binding constraint, and none of the above applies.

The failure mode we see most often is not choosing wrong on architecture. It is choosing on force rating alone and discovering that the material spread was wider than expected, so the machine that was comfortable on annealed copper is at its limit on hard temper. EN 13601 puts hard copper at 290 to 360 N/mm² tensile against 220 to 260 for soft annealed, which is a substantial change in the force a bend demands. Work out what the hardest material and thickest section you actually buy will require before you compare drivetrains, using the bending force calculator and your own worst-case bar rather than a nominal one.