What's new
What's new

Comparative Costs for Robots

atomarc

Diamond
Joined
Mar 16, 2009
Location
Eureka, CA
I have a customer that has a process that would lend itself to automation via a robot. After reading a thread on this forum I looked at Universal brand robots. The look neat and their programming seems relatively easy but their price caught me off guard...$48K for the medium size machine.

Is that price a deal...do other, larger units sell for hundreds of thousands of dollars. I looked on eBay and saw used robots for under $15K. I think the application I have in mind is for a small arm with light weights involved..almost a pick and place scenario.

So..would one expect to pay $50K + for a robot, software and accessories.

Stuart
 
A 3-6 kg robot should be in the $25-35k range for a one-off buyer. If you were buying 20 of them over the next 2 years or so, the price could decrease significantly.
 
A 3-6 kg robot should be in the $25-35k range for a one-off buyer. If you were buying 20 of them over the next 2 years or so, the price could decrease significantly.


So once you have the robot the software an integration cost are what?
Will this swamp the price of the hardware?
Will the I/O fold nicely to my machine or cell. How do I handle oh-poops......
 
Well, maybe 50K for just a cheap robot, no accessories. When I was doing robotics (way back) they were much more expensive - for a single robotic arm, you could figure on spending at least $150-250K for a medium load 5 axis cell fully built. But then again, you would also have nearly 20K just in a PLC to integrate the thing. It also comes down to precision and repeatability, how accurate do you need it to be? High speed pick and place robots that can repeat to with within thousandths at high speed can be much more. A used robotic arm that was intended for waving around a paint gun can be had for a few thousand. Used robots may be found quite cheap, but you run into worn joints and sloppy drive systems, meaning they are not going to be as accurate or as repeatable. And they will likely need maintenance, which can be quite specialized and expensive. The real issue is the control system; it's almost all proprietary, all expensive, and mostly undocumented. An older robotic control cabinet can be worse than an old CNC mill, you have no idea what they might have done inside that thing to integrate it into their assembly line, and to maintain it. And if the control isn't working, that robot is just going to sit there laughing at you. Those ebay robots are cheap for a reason. Remember, you need the robot itself, a controller, the end effector( the bit the does the work or holds the part - does not come with the robot. This also may need to have some compliance to account for positioning errors), outboard power for your end effector (electricity, air, vacuum, whatever you are doing), control circuitry to integrate into the rest of your work cell (usually a PLC), safety guarding (light curtains, lasers, moving walls or automatic doors, etc.), an emergency stop circuit, and workholding on the incoming and outgoing sides of the workflow.
 
Lots of stuff I didn't think about. Looks like this may be a wet dream unless the customer wins the lottery. Those 'Universal' arms sure look cool, especially for a programming dummy like me, and no cage or light curtain required, according to the manufacturer...don't know about OSHA or Cal-OSHA though.

Thanks for the information.

Stuart
 
As was said...
A typical industrial robot 3-5 kg, is 40-60k$.

Working, for you, is == 100-130k$ delivered.
Grippers, integration, crapola they add-on.

I may be working on a very much cheaper solution, in the near term...
But the initial versions FY 2017 wont be polished/shiny/packaged.
 
We integrate robots for grinding / polishing applications - cycle times of 6 - 8 seconds per part and required to operate 3 shifts 6 days a week.

If you pay $xx,000 for a robot, our rule of thumb is that by the time you automate feeding it parts, creating EOAT to handle the parts (or families of parts), putting it in a cage, adding safety logic, providing a user friendly HMI, etc. etc. . . . your end price will be closer to 3x $xx,000 or more.

If you are ordering multiple cells, the price might get as low as 2x . . . if you have inline gauging, QA processes, the price might be 4x or higher.
 
For a standard little guy, figure 25k for a basic one and 35k for waterproof or clean or extended reach. That includes the controller and the software. Not the grippers, end effectors, mounting, caging, vision and feeding.

Less if you can use a SCARA instead of a 6 axis.

More for collaborative, but I am not sold on them quite yet. Give it five years and there won't be any other kind for small robots, though. Universal charge about a 50% premium right now, and you should have a chat with your local OSHA before you buy one.
 
Depending on how basic your pick and place is, do you need a robot? Can you get by without full positional feedback, reprogrammability, range of motion? In other words: is this something you can accomplish with a couple of air cylinders and a nice complaint gripper?
 
We integrate robots for grinding / polishing applications - cycle times of 6 - 8 seconds per part and required to operate 3 shifts 6 days a week.

If you pay $xx,000 for a robot, our rule of thumb is that by the time you automate feeding it parts, creating EOAT to handle the parts (or families of parts), putting it in a cage, adding safety logic, providing a user friendly HMI, etc. etc. . . . your end price will be closer to 3x $xx,000 or more.

If you are ordering multiple cells, the price might get as low as 2x . . . if you have inline gauging, QA processes, the price might be 4x or higher.

Interesting.
When I was in the hardware had a higher impact on final price.
Has the hardware dropped...relatively to labor.
Good to know.
 
Interesting.
When I was in the hardware had a higher impact on final price.
Has the hardware dropped...relatively to labor.
Good to know.

I have never seen the hard side cost more than the integration and soft side but I've only been doing it since 1977 so perhaps it was different back a while.
It does depend on what you consider total cost. Do you count sizing/selection, prints and documentation?
Some places toss this "overhead" into markup so the base price is driven by hardware.
Bob
 
We put together a machine tending cell for a milling machine. Our first experience with such a project.

The cost were:
Used and refurbished Fanuc M6ib robot from authorized Fanuc integrator (15K)
We had them make a couple of sample EOAT tools that we could use as and example for building our own (10K)
Shipping/Rigging/Travel to vendor for some basic training (4K)

After this starting point, I spent about 3-4 months off/on just thinking how I would handle our high mix /low volume product mix and integrating the robot into a cell (Remember I had zero experience with this stuff). I even went to IMTS to look around at some of the sample automation cells on display.

Then the fun really started and we built ourselves:
1.) Cage area - Industrial fencing found on craiglist + hardware (<$300)
2.) Material area (approx $1000 for shelving - surplus 80/20)
-Material area contains shelf space that holds aluminum fixture plates (set up for 30 plates each 6" x 4")
3.) Fixture Plates (approx $2000 for qty 30)
-Each fixture plate must be loaded with specific part(s)/operation and has a two pin interface and boss that the EOAT can grab on to. This way we can use the same EOAT for all fixtures.
4.) New EOAT - Built an air activated collet system that picks the fixture plate (Approx $1200)
5.) Hydraulic Vise ($1700) - Kurt 6"
6.) Air over hydraulic pump and sensors to run vise ($500 for pump/fittings/mounting stand/sensors/hoses)
-We got away here cheap as this can cost a lot. Ebay has loads of industrial surplus for pennies on the dollar
-The vise holds all fixtures flat and the absolute position doesn't matter too much as each fixture has locating
elements so we can pick off positions with the milling machine probe.
7.) 'Servo Door' - Built a pneumatically activated door. Surplus from Ebay actuator and solenoids and sensing elements (Approx $700)

I spent about a month on/off (say a good 5-6 solid days) learning about the input/output options on the milling machine, experimenting with signals, and thinking how I would arrange/wire everything.

I came up with a system where each individual element is under the control of an individual PLC. The milling machine has spare M codes that I assign to the various functions. There is an 'MFIN' signal that can be used to tell the CNC that the operation is finished.

For example M25 will Open Door. M25 signal is wired to the PLC that handles only the door. The PLC will operate the door via air solenoid. When the door is fully open, the sensor on the door actuator will tell the PLC that "The door is open". The PLC will then signal the MFIN to the CNC. I use small/cheap PLC from a company called Velocio. They are cheap enough so that you can use one for each element and the software is free.

PLCs for DOOR, VISE, CAGE, ROBOT (approx $200). Each element has inputs telling it to do something with and MFIN
output telling the CNC it is done. The CNC will never move on to the next step if it does not get an MFIN signal.

The method we used keeps the elements seperate is a bit more managable then some of the more complicated PLC setups I have seen as it makes it more complicated to "seperate everyhting". Since we need only a 6" hydraulic vise for all workholding, it is easy and quick to setup.

So we were all in for around 45K, which is about 1/3 the cost of a ready made cell. Not that this was a bargain as I spent a lot of time figuring everything out. I only did it this way as I did not have an immediate project that needed to hit the ground running, but more of an academic exercise with eventual financial benefit.
 
With our volume pricing, I can put a robot in for a pick and place for the same money or less than I can build a dedicated pick and place. The robot has other uses after it's life as a pick and place, the dedicated pick and place would most likely be scrapped.

We can buy a 10kg robot and fully integrate it for typically less than $60k for a machine tender. For other automation (packaging or unpacking or other things, the cost may run as high as $120k depending on how many cameras and other peripheral costs.

BUT.......

We are our own MTB for these things, with EE's and ME's on staff and thus the project is considered non-profit - just costs for materials and labor (labor rate includes overhead costs). Any outside firm you use is going to have to charge SG&A and Profit and thus there will be a significant cost increase.
 
With our volume pricing, I can put a robot in for a pick and place for the same money or less than I can build a dedicated pick and place. The robot has other uses after it's life as a pick and place, the dedicated pick and place would most likely be scrapped.

We can buy a 10kg robot and fully integrate it for typically less than $60k for a machine tender. For other automation (packaging or unpacking or other things, the cost may run as high as $120k depending on how many cameras and other peripheral costs.

BUT.......

We are our own MTB for these things, with EE's and ME's on staff and thus the project is considered non-profit - just costs for materials and labor (labor rate includes overhead costs). Any outside firm you use is going to have to charge SG&A and Profit and thus there will be a significant cost increase.


Are the EE's, ME's and IT staff's costs directly carried to these projects including the meeting/planing times?
How do you track the design/engineering cost, is it fully allocated over a year between jobs?
I assume the overhead handles the manager's salary, office space, basic IT support and such.
I have seen big companies make a black hole here, particularly you see it in engineering departments.
You may be as good and as fast inside as any outside integrator like Motion but few places are.
This speed and knowledge in a area of concern sometimes makes up for the markup.
Can you really be a world class, faster than those who specialize in it, builders in robots, machine vision, etc.
Can your volume hire the best and brightest in the country?
For sure as a engineer you want to do this yourself inside and a huge drive here.
In the end are you a MTB or a maker of products? Why not build your own machine tools and code your own controller?
Why not build your own operating systems for your office systems?
Bob
 
Are the EE's, ME's and IT staff's costs directly carried to these projects including the meeting/planing times?
How do you track the design/engineering cost, is it fully allocated over a year between jobs?
I assume the overhead handles the manager's salary, office space, basic IT support and such.
I have seen big companies make a black hole here, particularly you see it in engineering departments.
You may be as good and as fast inside as any outside integrator like Motion but few places are.
This speed and knowledge in a area of concern sometimes makes up for the markup.
Can you really be a world class, faster than those who specialize in it, builders in robots, machine vision, etc.
Can your volume hire the best and brightest in the country?
For sure as a engineer you want to do this yourself inside and a huge drive here.
In the end are you a MTB or a maker of products? Why not build your own machine tools and code your own controller?
Why not build your own operating systems for your office systems?
Bob

We are set up as mini-plants within the facility. Each department has it's own engineering group to manage production engineering and department projects. We also have a division within the corporation that does machine tool and automation development and projects. We do build our own machines, rebuild machines including control systems, or heavily modify store-bought machines to meet our needs and we do code our own controllers, robots (programming, not operating software), machine tools and vision systems.

We have some of the brightest people I've ever been around (and of course, some of the not so brightest too.. but those generally don't last long). Our Engineering folks are mostly exceptional. Almost all are "Type A" personalities. You retain those types by allowing them to do what they really want to do....and that is design and build things that are cutting edge or work on leading edge things that challenge them. (For instance, we had vision-guided robotic bin picking in place and operating in production 4 years before it ever showed up at IMTS. - not saying it hadn't been done, but there wasn't a commercial solution available when it was developed in house.)

All of the costs for the project are carried by the individual project. We have many projects going on at any one time. They actually clock in to the project, this includes planning meetings and the like. All costs are tracked. The rate they are charged out covers all overhead for that department including salaries & benefits, floor space, equipment, management, etc. They are treated as a separate entity accounting wise. It's really not any different than hiring an outside company, except they work exclusively for us or other facilities within the corporation and they don't make a profit. The profit is realized in the cost reductions or throughput improvements for the projects they do for the production areas.

So for your last one........ we are a maker of products with an in-house MTB that works to help us reduce our operating and manufacturing costs for the products we make and to provide innovation to allow us to market more desirable products with technological advantages over our competition.
 
I have never seen the hard side cost more than the integration and soft side but I've only been doing it since 1977 so perhaps it was different back a while.
It does depend on what you consider total cost. Do you count sizing/selection, prints and documentation?
Some places toss this "overhead" into markup so the base price is driven by hardware.
Bob

I was referring to the relative impact of hard ware vs software.
Mg says 2-4 x ratio.
I seem to remember 1-3x a decade ago.
Of course now I'm back on the loop...we'll see what I learn.
 
We put together a machine tending cell for a milling machine. Our first experience with such a project.

The cost were:
Used and refurbished Fanuc M6ib robot from authorized Fanuc integrator (15K)
We had them make a couple of sample EOAT tools that we could use as and example for building our own (10K)
Shipping/Rigging/Travel to vendor for some basic training (4K)

After this starting point, I spent about 3-4 months off/on just thinking how I would handle our high mix /low volume product mix and integrating the robot into a cell (Remember I had zero experience with this stuff). I even went to IMTS to look around at some of the sample automation cells on display.

Then the fun really started and we built ourselves:
1.) Cage area - Industrial fencing found on craiglist + hardware (<$300)
2.) Material area (approx $1000 for shelving - surplus 80/20)
-Material area contains shelf space that holds aluminum fixture plates (set up for 30 plates each 6" x 4")
3.) Fixture Plates (approx $2000 for qty 30)
-Each fixture plate must be loaded with specific part(s)/operation and has a two pin interface and boss that the EOAT can grab on to. This way we can use the same EOAT for all fixtures.
4.) New EOAT - Built an air activated collet system that picks the fixture plate (Approx $1200)
5.) Hydraulic Vise ($1700) - Kurt 6"
6.) Air over hydraulic pump and sensors to run vise ($500 for pump/fittings/mounting stand/sensors/hoses)
-We got away here cheap as this can cost a lot. Ebay has loads of industrial surplus for pennies on the dollar
-The vise holds all fixtures flat and the absolute position doesn't matter too much as each fixture has locating
elements so we can pick off positions with the milling machine probe.
7.) 'Servo Door' - Built a pneumatically activated door. Surplus from Ebay actuator and solenoids and sensing elements (Approx $700)

I spent about a month on/off (say a good 5-6 solid days) learning about the input/output options on the milling machine, experimenting with signals, and thinking how I would arrange/wire everything.

I came up with a system where each individual element is under the control of an individual PLC. The milling machine has spare M codes that I assign to the various functions. There is an 'MFIN' signal that can be used to tell the CNC that the operation is finished.

For example M25 will Open Door. M25 signal is wired to the PLC that handles only the door. The PLC will operate the door via air solenoid. When the door is fully open, the sensor on the door actuator will tell the PLC that "The door is open". The PLC will then signal the MFIN to the CNC. I use small/cheap PLC from a company called Velocio. They are cheap enough so that you can use one for each element and the software is free.

PLCs for DOOR, VISE, CAGE, ROBOT (approx $200). Each element has inputs telling it to do something with and MFIN
output telling the CNC it is done. The CNC will never move on to the next step if it does not get an MFIN signal.

The method we used keeps the elements seperate is a bit more managable then some of the more complicated PLC setups I have seen as it makes it more complicated to "seperate everyhting". Since we need only a 6" hydraulic vise for all workholding, it is easy and quick to setup.

So we were all in for around 45K, which is about 1/3 the cost of a ready made cell. Not that this was a bargain as I spent a lot of time figuring everything out. I only did it this way as I did not have an immediate project that needed to hit the ground running, but more of an academic exercise with eventual financial benefit.

This was a great post - thanks for the detailed info.

Brent
 








 
Back
Top