Most robot purchases in India go wrong before the purchase order. A payload figure is copied from a brochure, reach is chosen by eye, and the cell is designed around a machine that cannot do the job.
This guide gives you a decision order: define the task, size the robot, check the cell, then compare the commercial options. It is written for a factory owner or engineer buying a first or second robot, not for a systems integrator.
Start from the task, not the robot
Write down what the robot must physically do, in numbers. Parts per hour, part weight, part size, how it arrives, where it must go, and what accuracy the process needs. Every robot decision follows from that list, and most disputes later trace back to a list that was never written.
| Question | Why it decides the robot |
|---|---|
| What is the heaviest part, including gripper? | Payload rating, with margin |
| How far must the arm reach? | Reach envelope and mounting position |
| How accurate must the place be? | Repeatability, vision, fixturing |
| How many parts per hour? | Cycle time and number of robots |
| What does the part look like? | Gripper type and sensor choice |
| What environment? | IP rating, cleanroom, temperature |
| Who operates it? | Programming interface and safety |
Size the payload with margin
Rated payload is measured at the wrist flange, with the load centred. A real gripper is not a point mass: it has length, so its weight creates a moment. A 3 kg gripper hanging 200 mm from the flange loads the wrist more than a 5 kg block bolted flat against it.
- Add part weight, gripper weight, cabling and any tooling.
- Add 20 to 30 percent margin for acceleration and for future parts.
- Check the payload curve, not just the headline number; capacity falls as reach extends.
- Keep the centre of gravity close to the flange.
- Confirm cycle time at that payload, not at the brochure’s maximum speed.
Reach is a shape, not a number
A robot with 1.3 m reach does not cover a 2.6 m sphere in practice. The useful envelope excludes the area behind the robot, the areas near singularities, and the space consumed by the gripper and part. Mounting height changes everything.
Take the CAD model of the cell and put the robot in it before committing. Most robot makers provide CAD and simulation software for free, and an hour spent moving the base in simulation saves a week of mechanical rework during commissioning.
Repeatability is not accuracy
Repeatability is how tightly the robot returns to the same taught point. Accuracy is how close it gets to a coordinate you calculated. A robot can be very repeatable and poorly accurate, which is fine for pick and place from fixtures and not fine for offline-programmed welding.
| Task | What matters | Typical approach |
|---|---|---|
| Pick from a fixture | Repeatability | Teach the point |
| Pick from a conveyor | Repeatability plus sensors | Teach, then correct with vision |
| Arc welding | Accuracy and path | Calibration plus fixtures |
| Assembly with clearance | Repeatability and compliance | Force sensing or floating holder |
| Inspection | Accuracy and stability | Calibrated cell, temperature control |
Cycle time arithmetic before the demo
Cycle time is path length divided by speed, plus acceleration and settling, plus gripper and handshake time. A robot that moves fast in a demo can still miss the target once the gripper waits for a signal.
cycle_s = approach + grip + transfer + place + release + clearance
# example: 0.6 + 0.4 + 1.2 + 0.5 + 0.3 + 0.4 = 3.4 s
# 3600 / 3.4 = 1058 parts per hour at 100 percent efficiency
# plan at 75 to 85 percent for a real lineMeasure the gripper time with the actual gripper. Vacuum cups need time to build vacuum; a two-finger gripper needs time to close and confirm. These are often the largest single term in a short cycle.
The cell around the robot
- Part presentation: fixture, conveyor, bin or tray. Bins need vision or a bowl feeder.
- Guarding and light curtains, with a safety relay rather than a standard PLC input.
- Controls: who owns the sequence, the robot or a PLC.
- Services: compressed air, power, network and the panel location.
- Operator access for loading, unloading and clearing jams.
- Maintenance access, including space to replace a motor or a cable.
Budget the integration separately from the robot. In Indian projects, a robot that costs fifteen lakh can easily need another eight to fifteen lakh in gripper, fixtures, guarding, conveyor, panel, installation and programming. Quotes that only mention the robot are incomplete.
New, used or cobot
| Option | Strengths | Watch out for |
|---|---|---|
| New industrial robot | Warranty, support, spares, known history | Highest capital cost |
| Used industrial robot | Much lower price, proven models | Controller age, missing cables, no support |
| Cobot | Easier deployment, no fence in some cases | Lower speed, payload and reach |
| SCARA or delta | Fast for planar tasks | Limited workspace shape |
A used robot can be an excellent first purchase if the controller is supported, the manuals exist and the mechanical condition is verified. A used robot with an obsolete controller and no spares is a project, not a bargain.
Questions to ask every supplier
Ask for separate dates for availability, dispatch, integration and production acceptance. The industrial robot lead-time planning guide shows how to check dependencies rather than treating the robot shipping date as the factory start date.
- What is included in the price: robot, controller, cables, software, training, commissioning?
- What is the local service response time and where are spares held?
- Can you show a live reference cell within driving distance?
- What is the warranty period and what voids it?
- Which safety functions are built in and what certification do they carry?
- What is the lead time, and how much of it is shipping?
- What training do operators and maintenance staff receive?
A supplier who cannot answer the service and spares questions is selling a machine, not a working cell. In a factory, uptime is the product.
Related: industrial robot cost in India, buying a used industrial robot, cobot versus industrial robot, payload calculation, conveyor integration.
What payload margin should I allow when choosing a robot?
Plan for 20 to 30 percent above the total moving mass, which includes the part, gripper, cabling and any tooling. Also check the payload curve at your working reach, because capacity falls as the arm extends.
Is a used industrial robot a good first purchase?
It can be, when the controller is still supported, manuals and cables are available, and the mechanical condition is verified by someone experienced. A used robot with an obsolete controller and no spare parts is usually a false economy.
How much does the cell around the robot cost?
In Indian projects, integration commonly adds 50 to 100 percent of the robot price once you include the gripper, fixtures, guarding, conveyor, panel, installation and programming. Always ask for a line-item breakdown.
Cobot or industrial robot for a small factory?
A cobot suits low volume, frequent changeovers and shared workspace. An industrial robot wins on speed, payload and reach for a fixed high-volume task. The decision is about the task and the operators, not about which is newer.
How do I confirm cycle time before buying?
Ask for a simulation or a demo with your part and gripper, and add your handshake and gripper times. Brochure speeds are measured under ideal conditions and rarely match a real cell.
Uptime evidence should shape the purchase, which is why the monitoring guide recommends measuring the existing machines first.
When two arms look identical on paper, the repeatability and accuracy guide tells you which figure your task depends on.
Reach must be checked against the fixture, not just the part, which the torch access section explains.
If your parts are formed on a press brake, the robot bending cell guide shows which of the three automation levels fits your volume.
Running cost belongs in selection, so the robot cell energy guide gives you the annual kWh figure to compare against.
If the work is long and straight rather than deep and complex, a gantry robot is usually cheaper than an arm on a track.
Painting is the one application where a general-purpose robot cannot be substituted, as the painting cell guide explains.
If the existing arm still suits the application, a retrofit may be the cheapest way to get a modern control system.
The quote you compare should be a complete cell, not a robot price; the integration cost guide explains the difference.
Before you shortlist a machine, confirm how you will fund it; the automation subsidy guide for MSMEs covers the main routes.
Sources and further reading
- ISO 10218-2 — robot systems and integration requirements.
- ISO 9283 — performance criteria and test methods for industrial robots.
- International Federation of Robotics — market data and application references.
