Start Production Faster with...

Robotic Machine Tending

How to improve CNC Machine Production

Time to Read: 34 min | Last Updated: April 2019

What is machine tending?

Definition of machine tending

Machine tending constitutes the largest use of collaborative robots on the market. Before going too deep into details, let's start with simple definitions.

The basic definition of tending is providing treatment for someone or something. In our case, machine tending means to load and/or unload a given machine with parts or material. Currently most machine-tending applications are done by humans. Modern machine shops often use CNC machines (lathe, milling, etc.). These machines must be tended by workers, who place raw material (typically known as raw or blank parts) in the machines and remove it once the machine has done its work. However, since qualified workers are becoming harder to find, companies are introducing robots into their workshops to make up for the lack of employees.

A robotic machine-tending process can be repeated endlessly, assuming the robot continually receives raw parts and the machine produces quality parts. Some industries use robots for a single step of production, like emptying injection molding machines or CNC machines. When production is running around the clock, robots enable you to minimize cycle time and run the process continuously by removing parts from the working area of the machine.

Read more:

Machine tending with a cobot

Collaborative robot, or cobot for short, is a general term for power- and force-limited robots. These are robots that can be used without machine guarding or that include other safety features. Cobots can be placed beside a machine or a person and set to perform a certain task without needing to be fenced off from the surrounding environment.

That being said, introducing a cobot around a CNC machine and programming it to load and unload parts is not as simple as it sounds. You will need to choose a robot that can accomplish the job and that can reach a level of performance comparable to a human worker.

What is involved with a robotic machine-tending cell?

A CNC cell is pretty complex. However, there are recurrent items that represent the core of a robotic machine tending cell. The anatomy is broken down in the following terminology.

Anatomy of a machine-tending cell


  1. CNC machine: Autonomously machines parts (CNC stands for computer numerical control).
  2. Vice: Holds parts while they are being machined.
  3. Controller: Coordinates the machine’s motions.
  4. Operator: Operates the machine (in a human-operated cell); operates the robot (in a robotic cell).
  5. Teach pendant: Used to program the robot.
  6. Parts: Go into the machine as raw parts—generally one at a time—and come out as machined parts.
  7. Gripper: Grasps the raw material, places it in the machine, and collects the machined part after the transformation process.
  8. Robot: Performs the tedious actions of loading and unloading parts for the operator, including opening and closing the CNC door.
  9. CNC door: Keeps metal debris confined within the CNC machine and prevents part projection while the machine is running.

Collaborative robots are used for machine tending in machine shops around the world because of their many advantages. In fact, this type of robot is safe to use around workers, can interact with the machine, and is easy to install. You may want to look at the following video to get a better idea of the different steps necessary to accomplish machine tending with a Universal Robot and a Robotiq 2-Finger 85 Gripper.

Read more:

Business examples

Companies are using cobots (collaborative robots) to accomplish three things: expand production, redirect human workers to higher-value tasks, and ensure everyone's safety. In this section, we'll look at several examples of how they're making it happen.

Expanding production


At Trelleborg Sealing Solutions, a Danish manufacturing company, collaborative robots are being used as smaller, safer alternatives to traditional industrial robots. Each cobot has effectively become an accessory to each CNC machine. As Jasper Riis, the company's head of production, says, “When we buy a new machine we also order a robot for it.” This is a completely valid approach, and it is clearly very effective for Trelleborg's purposes.

Machine operators in the company are able to program the robots at a basic level, but that seems to be all the robotics training they have at this stage. The operators' jobs have remained pretty similar to the machine tending they did previously. The main difference is one of scale. Before, each worker would tend three machines simultaneously; now, they tend eight.

Read more:

Lowercase NYC

To bring its eyewear to life, Lowercase NYC imports raw material in sheets. The sheets are cut down into small tablets, and then turned into glasses frames in a CNC machine. The machine is tended by a UR5 cobot from Universal Robots and a Robotiq Gripper. The robot picks up the raw material, loads it into the first position, closes the door, presses the start button, and then does it all again for the second position. After machining, the robot removes the finished pair of glasses and places it in a bucket. The eyewear then goes through many steps of fine-tuning before being ready to ship.

Each production batch in the CNC machine averages 500 units, with around 30 to 40 units of each style. But even with such a small volume, the co-founder of Lowercase NYC Brian Vallario says the automated process is by far the best solution. ”I only have a few manual tweaks to do on the vices every time. The rest is all automated,” he explains. “This automation is important for us. Eyewear production is a very labor-intensive process, and we are such a small team that any improvement we can make in our efficiency is huge. Having a product that allows me to sit at the computer and work on design or go work on the more labor-intensive stuff that can’t be done by machines is a big plus for us.”

Read more:

Redirecting human workers

Some people might read the stories above and think, “Sounds like the robot took that machinist’s place and now that guy is out of a job!” But that’s not true. The point of collaborative robots is to do more manual tasks and free up human workers for more valuable tasks. Instead of providing physical labor, workers can spend more time on quality control, part verification, process optimization, and making sure production is running smoothly. Tasks like equipment maintenance can be done more often, and all without having to worrying about whether the machine is being fed. Cobots enable employees to:

  • Focus on process optimization

  • Focus on product quality and verification

  • Be redirected to more important or pressing tasks

  • Work in a more stimulating environment

Instead of standing in front of the machine waiting for the next part, workers can keep learning new things. Ultimately, employees get more stimulating jobs and and managers get better production. It's a win-win.

Tegra Medical

Tegra Medical  started small when it introduced collaborative robots, beginning with a single UR5 robot before scaling up to three UR5s on different machines. A culture of training and innovation is a key part of Tegra's approach. As Hal Blenkhorn, director of manufacturing, says,: “No one is going to lose their job to a robot. We're trying to [add more value to] employees, to train them in new skills, whether it's a different operation or making them the robot supervisor in that area.”

The culture of in-house expertise let Tegra get innovative with its use of robots. Taking what they learned from their first three integrations, the team embarked on an ambitious application for their next robot, a UR10. As Blenkhorn explains, “The challenges in this last cell were running three different products on three different operations. It's unusual for us—and it's unusual in the industry—to have a mixed-model cell like that that's feeding three [different] products simultaneously.”

By taking over basic programming functions, the manufacturing team has given Blenkhorn more time to develop new applications. The operators are even able to get involved in developing these new applications. This is clear when Senior Manufacturing Engineer Paul Quitzau talks about Tegra’s next application: “We are very excited. We’ve purchased our fourth robot and it's going to go into another area of our company where we perform laser marking.” As for how easy it was to install, Quitzau says “I was able to piece together the entire application without much issue, and I'm not a programmer, I'm a mechanical person.”

Read more:


One reason why it is easy to use collaborative robots in machine-tending applications is their built-in safety features. Not only will the robot automatically stop when it comes in contact with something (or someone), but it can also be limited to moving in a certain space. It only takes a few minutes to configure a safety plan that will restrain the robot’s arm movements to a defined workspace. This reduces the danger of using a robot, and also reduces the need to buy fencing or other safety devices to prevent collisions with workers.

If people haven’t seen a collaborative robot at work before, they might be nervous about the idea of being so close to one. Demonstrating safety in action is a great way to relieve anxiety. Use your first project to showcase cobots’ safety features. Word will spread about how slowly the robot moves and how easily it stops, and people will relax.

Who determines cobot safety guidelines?

The most widely recognized source of safety guidance for cobots is the ISO/TS 15066. This is a technical report produced by the International Organization for Standardization (ISO). ISO/TS 15066 (Robots and robotic devices — Collaborative robots) is not actually a standard yet. It is a technical specification (TS), because more technical development is required before it is turned into a full standard.

For our purposes, this means that the safety limits it prescribes are likely to change. As we better understand the practicalities of working alongside a collaborative robot, the limits will be updated. However, even though it's not a full standard yet, we’ll still refer to it as a “standard.” You can find out all about the standard in our eBook ISO/TS 15066 Explained.

Read more:

Setting up a machine-tending robot

With your first robotic cell deployment, your primary objective should be getting a positive response to the project along with bottom-line return on investment numbers. Your first project is an opportunity to dispel myths and highlight benefits regarding safety, job security, programming ease, and other concerns. By implementing a robotic cell that’s safe, frees people up for other value-added work, and does what it’s supposed to do without requiring excessive programming expertise, you’ll gain enthusiastic support for future projects.

How do you set the first project up for success? Keep it simple. Our book Lean Robotics: A Guide to Making Robots Work in Your Factory advises, “if you need to choose between a simple, low-ROI application and a complex, high-ROI one, it’s best to go with the simple one… Far better to start simple, and make sure the first cell deployment is a success, so you can start creating value with it while building momentum for more ambitious future projects.”

For your first project:

  1. Implement a simple, proven solution.

  2. Get it up and running as quickly as possible with no unnecessary extras.

  3. Set a short timeline so you can quickly get back on track if the project fails.


Read more:


The design phase of a machine-tending application starts by looking at your actual process. Monitor your human-operated cell for a few days or weeks to get an idea of whether your investment in robotics will be worthwhile.

To track the current cell’s production, record the following data:

  • Production rate
  • Spindle time
  • Part defects

Next, record photos and videos of your current machine-tending process to see exactly what’s being done.

Read more:

Understand robot limitations

Keep reading for our cheat sheet on different robot specifications, which will help you choose the right robot.

Read more:

Robot reach

The robot’s reach, or work envelope, is the robot’s range of movement. If you want your robot to grasp parts, use a tool, and open a door, all these items must be located within the robot’s work envelope.

A machine-tending robot needs to take a blank part from somewhere (point A) and place it in the machine (point B). To calculate the minimum reach your machine-tending robot must have, measure the distance between these two points and divide it by two.


  • If you need a large reach, your robot's payload will necessarily be higher; a larger robot is a stronger robot.
  • A robot’s reach is also determined by its number of axes (degrees of freedom).

Robot payload

The robot’s payload is one of the most important specifications. Payload is the total weight the robot can carry. Since your robot will carry different types of parts, you will need to determine which parts will be the heaviest and select your robot consequently. To determine your required payload, sum the weights of your tool and the heaviest raw part you want to carry.

Calculate part weight by multiplying the part’s volume (W x L x H) by its density (g/mm^3). Or use CAD software to determine the weight of the part; that way you won’t need to do the math.

In order to have safe robot acceleration and speed settings, the robot should carry a weight no more than 90–95% of its maximum payload. In fact, carrying a load too close to the maximum payload can cause an error and stop your program.

Quick tip: Always add an extra 20% to your final payload calculation, to enable maximum accelerations and prevent errors while running the program.

Read more:

Robot repeatability

Although CNC machines are super repeatable and you want your parts to be precise, you don’t need your robot to be as precise as your machine. Machine-tending robots are typically used to do first setups where the part is usually larger than the finished part, which allows for small variations in positioning.

The robot should be repeatable within 0.1 mm. To ensure repeatability, use mechanical stoppers or a force-torque sensor.

Quick tip: Make sure to have a compliant component in your program in order to allow the vise to close and not be restrained by the robot. In other words, leave at least one axis “free” when inserting a part in a vice or lathe, to prevent wear on the robot actuators.

Robot end effectors (grippers)

Signatures-Grippers-RobotiqYour gripper is the robot’s “hand,” but it’s not nearly as versatile as your hand. Grippers work best when parts have at least two parallel surfaces. The stroke of the gripper will limit the range of parts you can handle.

Try to choose an adaptive gripper that can handle different shapes and, more important, different sizes of parts without modifying fingers or the robot’s programming.

Quick tip: Make sure your gripper can handle 85–95% of the parts. The rest can be loaded manually if need be.

Gripper payload

Like robot payload, gripper payload is the amount of weight the gripper can handle. Make sure to respect this payload to ensure gripper longevity.

Quick tip: In addition to respecting payload, make sure not to max out the gripper’s allowable torque. If the bulk of the object's weight is always grasped by the gripper’s fingertips, the gripper will wear prematurely.

Gripper dexterity

Grippers often have very limited dexterity. In fact, if your application requires the operator to handle several different parts at the same time and perform operations with both hands, proceed carefully. Robots can only do one thing at a time. This doesn’t mean you can’t use robots, but it does mean you will most likely have to redesign your process.

Here are some of the most important robot dexterity factors, along with questions to ask when defining your needs:

  • Object size. How big are the parts? Are they identical, or a variety of sizes? How does this compare with the reach required of the robot?
  • Object shape. What shape are the parts? Do they have many complex edges or a simple geometrical shape? Are they spherical or otherwise difficult to grasp?
  • Gripping strategy. How could the parts be grasped (e.g. with an encompassing grip, internal grip, or suction)? Are there different ways to grasp the same parts? Are the parts delicate and so require a particular gripping strategy?
  • Reach. How much does the robot have to “stretch” to reach all important locations in the workspace? Will the robot use its entire workspace, or just a small part of it? Does it need to approach locations from many different angles?
  • Speed. What cycle time is required for each action?

People sometimes think factors 1-3 only relate to the robot's gripper, and factors 4-5 only relate to the manipulator, but they’re actually all interrelated! One factor in isolation does not necessarily make a robot “dexterous.” You can only get a full picture of dexterity when you consider all the factors together.

Read more:

Gripper stroke

Your gripper selection will depend on the parts to be handled. If your parts vary in size, you will need a flexible gripper. Robotiq’s 2-Finger 85 Gripper can accommodate parts between 5 mm and 80 mm without any additional programming. If you will always be handling parts of the same size, you may want to go with a more rigid custom gripper that will perfectly fit your part.

Quick tip: A flexible gripper can grasp both raw and finished parts. A 2F85 Gripper will also allow you to adapt the force of the gripper in cases where the finished product is more fragile than the raw part.


Machine interface

In machine tending, the interface between the robot and the CNC machine is an important part of the integration. This is not always a given for different machines produced by different manufacturers. In order to optimize your cycle time, you should make sure the two machines can talk with each other so they know each other’s status. They will need to communicate statuses like CNC program done, door can be opened, vice closed, robot in motion, and part in place.

Quick tip: Make sure your robot and CNC or other machine can share the same interface. Some robot models may not speak the same language as your machine.

Ordering devices

Picking the raw part and placing the finished part must be done in a structured way. We recommend starting with a simple device that will position the part in the same spot every time. This is relatively easy to program and highly repeatable, especially since most cobots have built-in wizards to simplify path programming.

Quick tip: Keep it simple by using matrices or stacks to position parts in the same place every time.

Starting Point

In machine-tending applications most of the human operator’s manual process involves placing the part, tightening the vice, and starting the program—and doing all these tasks in reverse order once the part has been machined.

For the robot to place the part correctly, you'll need to use mechanical stops or other devices to guide the robot.

Read more:

CNC Cycle Start

To start the CNC machine, either have the robot push the start button (like a human would do) or use the PLC to send a signal to the CNC to automatically start the program. We don't recommend trying to coordinate this functionality with timers. It’s is a recipe for disaster: timers don't adjust over time, and don't respond to an exception, so you could end up with the robot butting against the CNC machine, trying to remove a part while the CNC door is still locked for milling.

Automated vise

The vise inside the machine must fit the machined parts precisely. Most vises are equipped with a pneumatic or hydraulic system that keep the part in place during the machining operations.

Since your robot will be loading and unloading parts on an unattended machine, you should use a vise or chuck that will open and close automatically. Most automated vises are air powered. This means you will need to use a small programmable logic controller (PLC) or the robot’s control box to send signals to the vise indicating when it’s time to open or close.

Actuated door

There will come a time in your integration where you will have to decide if you need an automated door or not on your CNC machine. If you are buying a brand new machine, take the option of an automated door. It’s generally worth the cost. If you are keeping your old machine, and want to improve your machine tending with a robotic arm, you could open and close the machine door with the robot instead of an automated signal. We’ve seen a lot of integration projects where the robot closes the CNC machine door itself. It’s slowlier, but it works.

Example of a robot-operated door


Example of an actuated door

Robot intelligence

Making decisions

Enthusiastic robotic newbies might like to say ‘’The robot will decide to do this or that according to the situation.’’ In reality, the robot will not decide anything; it will simply execute its programming. In other words, the decision will be made according to specific data.

Finding parts

Until recently, part recognition systems were really complicated. But with new built-in solutions like the Robotiq Wrist Camera and YuMi vision system, it’s quite easy to locate a part and do something with it.

Quick tip: Use a vision system in a simple application such as picking up a part lying on a flat surface. A more complex context, like the inside of a CNC machine, might make it too hard for the vision system to recognize the part. Keep it simple and situations with lots of variables (like multiple colors and reflections).

Communicating with the CNC

Many signals can be sent between the robot and the CNC, but the PLC can only process a limited number of I/Os. For that reason, it’s a good idea to limit the data that you would like to access. Prioritize the signals that are necessary for efficient machine-tending.

As explained in a discussion on DoF, you can easily use a simple signal from the CNC machine to start the robot program—such as a light signal—but it’s much harder to extract enough precise data from the CNC to trigger your robot. Keep the communications between the robot and the CNC machine as simple as possible. This will simplify your integration and keep you away from complex logic chains.

Quick tip: To trigger the start or end of a program, use discrete values like end position, a proximity sensor signal, or a light signal, not variables like distance, position, current, or force. Enforce strict barriers between “go” and “no-go” by leaving no room for interpretation.

Read more:


Redesigning your process

Product redesign

It’s safe to assume that most of the products you will be producing are for customers with very specific requirements. You probably can’t change a lot of features in the final product. However, you can still adapt your setup to the robot. If your gripper works best for grasping a certain part width, you may want to leave a feature of this width on the part at the first setup, or simply ensure it has two parallel surfaces to allow for a better grasp. Doing so will help the robot load the part for a second machine setup without any problems.

Watch how it worked for Scott Fetzer Electrical Group, at the end of the video.

Incoming raw parts

At human-operated machine-tending stations, we often see raw parts arrive all thrown together in a bin. This works fine for humans, but it’s more difficult for robots. Grasping is easier when there’s some space around each part, so make it easy for your robot by placing raw parts on simple trays.

There are two general ways to structure the parts that need to be machined. First is a classified racking system. In fact, if the parts are classified such that the robot knows exactly where the part is, then coordinates for the parts are easier to program. However, you might have to put some time and effort into the racking design. It’s no problem if the raw parts will remain the same for a long time, but if the production needs to be changed and a new raw part needs to be classified often, it’s a real puzzle to design a racking that can fit every raw material.

The second way to order your raw material is with vision. Raw parts can be placed in a conveyor, bin, or drawer, depending on their size. Since the parts are not positioned precisely, the robot program must be paired with a vision system. The system takes photos to find exactly where to pick up the part. If the robot needs to grasp parts while they’re moving (such as on a conveyor belt), you can set up a dynamic vision system by pairing it to an encoder on the conveyor belt. Here’s a video that shows how to order raw parts using a vision system.

Watch this video where incoming raw parts are ordered with a moving belt

Machined part output

When analyzing your existing CNC part output you may realize that it requires dexterous manipulations. Putting finished parts in a box or a bin is easy for a human, but it can be complex for a robot. Some of the best machine-tending cells we’ve seen have used trays. These simple devices maintain a lot of room between the parts and provide enough space for the robot’s fingers to open once the part is placed in the tray.

Remember, you don’t want to add another operation after the robot has unloaded a finished part from the machine. If the part is ready to go, place it in its final destination. A tray or box is ideal. Other good ideas: program the robot to leave a specific amount of space between the parts, so they don’t get damaged during transport.

Sensor management

Sensors are your robot’s eyes and ears. They’ll give you a lot of information, and you must manage all this data correctly if you want the robot to interact with other devices in your cell. Proximity switches and other physical sensors are commonly found on robotic machine-tending cells. If you can’t manage them with the remaining I/Os in your robot controller, you will need to use a PLC.
Also, even though collaborative robots have many safety features, they can benefit from additional safety sensors. You will most likely need a safety-specific PLC to manage all your different safety sensors.

Read more:


Prehension is probably the most complicated aspect of the machine-tending process. In fact, prehension includes the choice of an end-effector and robot. You need to know what kind of parts can possibly be handled in the cell. You also need to ask yourself if the cell will be dedicated to a single product or if the production will change every month for example. This criterion will decide the size of your component (Grip force and payload needed) and if you will need a flexible or rigid cell. In fact, you can use a two-position gripper for a production that will stay steady for years. On the other hand, you want to use a flexible robot gripper for production that changes relatively often. Grippers like our 2-Finger Adaptive Robot Gripper and even 3-Finger Adaptive Robot Gripper are ideal for flexible applications.

Error signals

In a human-operated cell, whenever a CNC error pops up on the controller screen, the operator will know to stop and check the machine. This is not the case for a robot-operated cell. It’s dangerous to leave the CNC machine running despite an error warning, so you will need to set up the proper reactions to any error message in your program.

Quick tip: Make sure you have a way to confirm that the part has been correctly grasped by the vise and gripper. We’ve seen a few cases where such conditions were not respected, and the machine was damaged due to an unclamped part flying around the CNC.

Murphy analysis (or process failure mode effects analysis (PFMEA))

The Murphy analysis may as well be called “the how not to crash my robot assessment.” There are many variables outside your control; a Murphy analysis helps you identify these potential problems and find solutions to them.

If you’ve noticed a potential problem but are convinced it will never happen, trust us: it will happen. Make sure to do a full Murphy analysis, also known as a PFMEA. This technique will allow you to go through all the machine and robot steps and prevent potential errors.

Read more:

Stable processes

Automating a process means that nobody will be watching the machines for a potentially long time. If the process is unstable, there will be many opportunities for problems to arise and all your production efforts could be wasted. This is not good for quality control. So to avoid a disappointing robot integration, make sure to bulletproof your process before you let it run unattended.
Treat your robot the way you would a new employee. Start by giving the robot a simple task. After a couple of hours of operation, you’ll inevitably find parts of the program that need changing. Make the corrections. Only after the program is solidly up and running should you strive for a new application or increase the operations in the same cell. Robots can’t do everything, so let them handle small tasks and leave the human operators to concentrate on complex ones.

Read more:

Machine-tending operation steps

Although there are many things to include when programming your robot, the following steps will form the pillar of your program.

Robot arm movement: All collaborative robots can be programmed by hand guiding. This is an easy way to program a robot since it doesn't require much experience in robotics or programming.
Gripper movement: This depends on which gripper you’re using. A Robotiq Gripper, for instance, can easily be programmed using a special embedded feature in the Universal Robots teach pendant.

Raw part delivery and grasping: First determine how the parts will be picked, then set up an ordering method so the raw parts are easy to grasp.

Raw part placement: This can be challenging, because parts usually have to be placed quite precisely in the CNC machine to ensure that the CNC program will run normally. To ensure repeatable part positioning, you should either add a way to mechanically lock the part in place, or add force sensing to the robot gripper or tooling so it knows where to place the part.

Door closing and opening: The CNC machine must be kept closed for safety reasons. If you already have an automatic door, set up the robot/CNC machine interface. If not, program the robot to open and close the door itself.

Finished part grasping and delivery: After the robot grasps the finished part and removes it from the machine, it should place the part in a convenient location for transport. Make sure the finished parts won’t damage each other, or your final product quality will be affected.

Other stuff: After that you can make the robot do whatever crazy stuff you want it to do. Clean the table, push some buttons, whatever!

Read more:


Optimize downtime

Validate finished parts

Depending on the part’s complexity, a metrology device can be inserted into the cell. Whether a camera that is measuring certain diameters or even probes directly on the end effector, use a measuring method that verifies the critical dimensions. Your process needs to be verified, so choose a verification method that is adequate for your application. Metrology can be done afterward, but don’t wait too long to verify your process because the situation can change really fast. If, for example, a tool is broken in the machine magazine, you don’t want to have to redo the production.

Several techniques can be used to measure features on a part. You may want to use the tool inside the CNC machine to make sure the part has been finished correctly, or add an external device or robot tool. For instance (and as shown in the video below), a force-torque sensor can measure the external diameter of a part.


 A force-torque sensor can measure the external diameter of a part

Robots can perform a variety of sub-assembly tasks, such as fastening parts together or inserting sealant. Be aware of what your robot arm might be capable of doing to save you time in your production process.


Dirty parts can create surface treatment problems, so part cleaning is a great way to ensure consistent part quality.


Here is a machine tending cell using pneumatic cleaning to clean away the debris.


A robot cannot perform machine setups, but your employees can. Since they don’t have to tend machines anymore, they can reduce downtime by preparing machines for the next production run. This will increase spindle time and allow you to output more parts.

Read more:

Return on investment

How much does a machine-tending cell cost?

The “do-it-yourself” kit

You can build your own custom cell with a collaborative robot and an end effector. This is probably the cheapest option, with robots available for as little as $25,000 and inexpensive grippers at around $5000. Programming can be done by hand-guiding, so this setup is easy to integrate and quick to earn an ROI.

The “I don't want to spend time integrating” kit

Some robot manufacturers and robot integrators are offering pre-engineered machine tending cells. Based around either a collaborative robot or an industrial robot, the kit comes with everything you need for your machine-tending application. However, the price of these systems can be quite high due to all the accessories needed and the integration time: at least $50,000. Still the setup remains relatively small and the payback period can still be quite short.

The “all in one” kit

Now we’re talking about a major investment. If you want a setup that handles many steps for your CNC machine or carries huge parts, you’ll have to pay more. You’ll need an industrial robot, safety fencing, a dedicated space for the machine-tending cell, and someone to program and synchronize everything.

Some companies offer built-in programming kits that are easier to work with, but these still require a huge time investment, in terms of training workers to use the robot. These cells start at $100,000, and there’s basically no upper limit. You can introduce all the fancy accessories you could think of. This kind of cell must be set up by a qualified integrator.


Here’s an overview of the costs you can expect from adding a robot vs. other options.


Implementation cost

Incremental cost


Adding a CNC

About $250,000

Labor to tend machine

Best if existing machines are nearing end of life



Per part, per contract

Best if cash flow is an issue

Adding a robot

About $50,000,

including end effector

and integration

Labor to manage robot

Best for increasing per-machine productivity

Read more:

Payback period


One reason to use a robot is to raise production. Using a time-based ROI calculation makes this benefit easy for everyone to understand. It’s easier to understand earning back the investment if you describe it as “within six months” rather than “after 1,273 parts.”

Use the following data to calculate the payback period:

  • Average spindle time before introducing the robot
  • Production rate before introducing the robot
  • Profit on the finished parts
  • Hourly machine cost, including employee cost and machine cost
  • Robotic cell cost
  • Average spindle time after introducing the robot
  • Production rate after introducing the robot

To calculate the payback period, find the difference between spindle time before you added the robot, and after:

Payback period = Δ spindle time = (robotic cell cost / machine rate).

For example, if your robotic cell cost $50,000 and your operating costs $100/hour to run, the payback period will be as follows:

Δ spindle time = ($50,000 / $100 per hour) = 500 hours.

This means you will earn back your investment after your machine has performed performed 500 more hours than it would have normally. In other words, after your robotic cell has performed 500 more hours of machine-tending than it would have if you had kept it as a manual cell.

Now, let’s say you will run the machine autonomously during the night shift, which represents 40 hours per week:

Night shift payback period - Δ spindle time / hours worked per week - 500 hours  / (40 hours/week) - 12.5 weeks or 3.12 months.

You can also calculate the additional profit you will earn from producing extra parts (as a result of the increased spindle time):

New machine rate = old machine rate - (Δ production rate × profit per part).

For example, if your production rate goes from 10 to 12 parts per hour, and you earn a $5 profit per part, your new machine rate is as follows:

New machine rate = $100/hour - (2 parts per hour × $5 per part) = $90/hour.

Calculating the new machine rate makes your ROI even more obvious, although the payback period alone is usually enough to convince management to add a robot to the shop floor.

Connected robots

Connectivity can give you a better idea of how well the robot is working. You might be able to use a simple device or tablet that will tell you everything you need to know about your robot, including working time, number of parts produced, and a whole lot more. This is not only useful for preventive maintenance, but also for getting alerted when the robot has a problem or if it has finished its task.

Robotiq launched the first software for optimizing Universal Robots performance. Insights is a web application that takes robot cell deployment to another level, letting users find out how productive their robot was in any given period of time.

Read more:

Your next robot

Because return on investment is not about how much you spend, but how fast and accurately you can produce your parts, you need to keep this data accurate and easy to use. If you present all this data to your manager, you might find that you are be able to have a couple of extra production boxes that will provide enough of an ROI to allow you to order another robotic cell.

Here are some rules of thumb when it comes to payback periods:

  • If the project can be reimbursed within three months, it’s a no-brainer
  • Within six months is generally a “go”
  • If it’s a year, you may need to push a little harder and provide further explanation
  • At two or more years, it might be hard to convince your manager (but not impossible)

Of course, if your first robot project succeeds with a quick ROI, you’ll have a great argument for why management should consider another robot for your shop floor.

Read more:

Let’s start production faster

Manufacturers work with Robotiq to take control of their robotics projects and improve productivity and quality. They choose our flexible, Plug + Play Components because they can be deployed easily across many stations. Our community of experts empowers manufacturing engineers to quickly deploy their robotic cells and build their automation skills.

Robotiq's tools and know-how simplify collaborative robot applications, so factories can start production faster


Jump To Section:

Download an offline copy:


Download the full bundle