Skip to content
TR Get in Touch

Our Solution Areas

Robotic Systems

Maximum production efficiency through industrial robot applications.

A robot on its own does not do the work; the cell does. Cycle time becomes predictable only when the arm, the gripper, the part presentation, the safety perimeter and the PLC that triggers the robot behave as one system. Our first decisions in a cell are about reach envelope, how the part is presented and where the safety boundaries fall — the robot brand comes after that.

Where it pays off

A robot investment returns where the work is repetitive, positionally defined and tiring for a person. Repeating the same motion thousands of times a day produces drift in a human as fatigue sets in; it does not in a robot. Not every manual task suits a robot, though: if it is unclear where the part will be picked from, or the product changes every week, the payback depends far more on the feeding arrangement than on the cell itself.

The cells we build most often:

  • Pick and place. Taking a part from a conveyor or feeder and placing it into the next station, a mould or a carton.
  • Palletising and depalletising. Defining patterns for sacks, boxes and crates; slip sheet placement, and a two-station arrangement so a pallet change does not stop the line.
  • Machine tending. Loading and unloading a CNC, press, injection moulding machine or test station in sync with its door, ready and alarm signals.
  • Welding, gluing and dispensing. Running the path at constant speed, matching corner deceleration to the flow rate, and avoiding excess material at start and end points.

How we work

We start a cell with a cycle time budget. Total time is broken into line items: robot motion, gripper grip and release, vacuum build-up, the machine’s door and ready signals, camera evaluation. Selecting a robot without that table usually buys a fast enough arm whose cycle is spent waiting.

Next comes layout and reach verification. The cell is modelled in offline programming (RoboDK, RobotStudio, Roboguide), so before anyone goes on site it is visible that the robot reaches every target with an acceptable axis posture, does not run into a singularity and does not collide with surrounding equipment. Simulation also keeps the layout negotiable while it still can be: if reach is marginal, we find that out before the robot is purchased.

Gripper selection is often the decisive engineering call. Vacuum, pneumatic fingers, magnetic or a purpose-built end-of-arm tool — the part surface, its centre of gravity and the accumulated tolerances decide. When payload is calculated without the tool’s own mass and the distance of the part from the flange (the moment of inertia), the robot derates its own speed and the cycle budget quietly slips.

For robot-to-PLC integration the signal table is written up front. The PLC runs the cell logic, the robot runs motion; if the handshake between them is loosely defined, recovering from a fault needs manual intervention every single time. Our test is this: after a fault, the cell must be able to return to its home position without losing the part and without sending an operator inside.

What we watch for

Calibration is done once and recorded. If tool and base frame values are not written down, the cell has to be retaught after a gripper change or a crash. Where a camera is used, hand-eye calibration goes into the same record.

Safety comes out of the stopping distance. The position of a light curtain or area scanner follows from the robot’s stopping time and the human approach speed; it is a calculated distance, not a preference for “as close to the wall as possible”. In collaborative operation the assessment is made on the application, not on the arm.

Part presentation is the real bottleneck. When the robot works correctly and the cell is still slow, the cause is usually parts arriving inconsistently. The choice between a vibratory feeder, a staged conveyor and vision-assisted picking sets the largest line item in the cycle budget.

Commissioning begins at reduced speed. A program is never first run at full rate; the path is verified step by step and speed is opened up gradually. That discipline is what stops a typo from driving the gripper into a mould.

How this connects to our other work

A robot cell never stands alone. Cell logic and line integration are built on the industrial automation side, while the part that determines a component’s position and conformity is designed together with machine vision and camera systems. The cell’s panel, safety circuit and field wiring come out of electrical and electronics work.

Frequently asked questions

Can you commit to a cycle time at the quotation stage?

Not a guarantee, but a measurable budget. Robot motion time can be predicted quite well in offline simulation, yet the real cycle also contains gripper open and close time, vacuum build-up, guard door travel and the machine's ready signal. So we break the cycle into line items, set a target for each, then measure them at commissioning and show the difference. Presenting a single number out of a simulation as a guarantee hides the fact that most of the delay usually sits outside the robot.

If we use a collaborative robot, do we still need a fence?

A collaborative robot does not make the application collaborative. The assessment applies to the application, not the arm: the mass of the transported part, its edges, the speed and the likely contact point decide the outcome. A cobot carrying a sharp-edged part or running at high speed still needs a fence or speed-and-separation monitoring. The force and pressure limits in ISO/TS 15066 frame that calculation, and nobody should declare "no fence needed" before the measurement exists.

What has to change to add a robot to our existing line?

Usually the biggest change is not the robot but the way the part is presented to it. A human finds a part in a mixed bin and orients it without thinking; for a robot the part position must either be fixed mechanically or determined by a camera. The second topic is the safety area: if the volume the robot sweeps crosses an operator walkway, the layout is redrawn first. We do not select a robot before both are settled.

Can we edit the program ourselves, or do we depend on you for every change?

We build the cell so that adding a new product is an operator task: positions and recipe data are separated from the body of the program and kept in editable tables. A new pallet pattern or a new part height can then be defined without writing code. Changing the motion logic does require working in the robot language; we train the team for that and the source program is delivered with its documentation.

Related Solution Areas

Smart solutions, secure tomorrows

Let us carry your production into the future

Tell us about the bottleneck on your line and we will come back with a measurable improvement plan. Write to us for an initial discussion and requirement analysis.

Get in Touch info@evasmartai.com