Thread: Any good/bad cobot stories?

Originally Posted by CarbideBob

A lower power and sensitive robot that can be used next to humans without the special guarding.
Friendly if they bump you or you bump them.
Bob

Thanks to all for answering!

Bob, so this would be something like they show in the hbo special* about robots, later in the doc they show it 'sensing' some interference and stopping?

*Ironically the title of the show is something like robots killing humans, I did not watch all of it, but did see the part about the sensitive robots (cobots?)...

So could a regular robot not be equipped with sensors that would allow it to move at full speed/power but stop if it hit something? Like a protected move on a probe?

My worst cobot story- we bought a UR10 several months ago, and we have been so damn busy busting out parts that we have had no time to touch it.

Originally Posted by Gobo

My worst cobot story- we bought a UR10 several months ago, and we have been so damn busy busting out parts that we have had no time to touch it.

Gotta ask, bragging or complaining?! My view: good problem to have!

Originally Posted by The Dude

Gotta ask, bragging or complaining?! My view: good problem to have!

Maybe a bit of both. We also just got in two Mori NLX's with all the goodies. It is an awful lot to absorb at once.

FYI, there's a "coffee with cobots" gig tomorrow in Portland. Sorry for short notice but if someone local is interested, PM me and I'll shoot you info on it. Basically a demo as I understand and just from Rethink Robotics (i.e. Sawyer).

The Dude

Originally Posted by The Dude

FYI, there's a "coffee with cobots" gig tomorrow in Portland. Sorry for short notice but if someone local is interested, PM me and I'll shoot you info on it. Basically a demo as I understand and just from Rethink Robotics (i.e. Sawyer).

The Dude

Sounds cool. Im on the other side of the mountains, otherwise I might have shown up.

Cool, never heard about it before.

Interesting that you mention Baxter as I program and use his predecessor Sawyer everyday!! So in the hands of "the guys upstairs" it was not good... No motivation and the setup was hacked. I'm a cnc mill programmer and setup guy and once we took it over it has been great! Couple things I have noticed. They have very bad accuracy. Full speed to get close to the part then VERY slow to get closer. Our hydro vise opens about .100" and we need every bit of it of that tells you something. BUT with all of that being said, with high quantity jobs it for sure pays. For an example we just ran a job that would take 3 hours to setup and about 60 hours of man hours in front of a machine. Well I set it up with the robot and have 6 hours of loading the "matrix"(table with parts that the robot picks from) and the job is done. Cost per hour on that mill is less than half of what the others are. So short version, it pays off of you have the proper setup and quantities to do it right. Hope this helps!

Originally Posted by CNCrobo

Interesting that you mention Baxter as I program and use his predecessor Sawyer everyday!! So in the hands of "the guys upstairs" it was not good... No motivation and the setup was hacked. I'm a cnc mill programmer and setup guy and once we took it over it has been great! Couple things I have noticed. They have very bad accuracy. Full speed to get close to the part then VERY slow to get closer. Our hydro vise opens about .100" and we need every bit of it of that tells you something. BUT with all of that being said, with high quantity jobs it for sure pays. For an example we just ran a job that would take 3 hours to setup and about 60 hours of man hours in front of a machine. Well I set it up with the robot and have 6 hours of loading the "matrix"(table with parts that the robot picks from) and the job is done. Cost per hour on that mill is less than half of what the others are. So short version, it pays off of you have the proper setup and quantities to do it right. Hope this helps!

After doing research. We feel the cobots are a waste of money.

We purchased a fanucmate today for one of the new Mills. Lots more features on the fanuc. And for $21,000 to boot.

Originally Posted by as9100d

After doing research. We feel the cobots are a waste of money.

We purchased a fanucmate today for one of the new Mills. Lots more features on the fanuc. And for $21,000 to boot.

Oh for sure!! Don't get me wrong that's next on the list! The collaborative part isn't with the speed and accuracy loss. But first I'm trying to figure out how to feed the bot with a saw... Then I'll start bugging them about a Fanuc lol

Originally Posted by CNCrobo

Interesting that you mention Baxter as I program and use his predecessor Sawyer everyday!! So in the hands of "the guys upstairs" it was not good... No motivation and the setup was hacked. I'm a cnc mill programmer and setup guy and once we took it over it has been great! Couple things I have noticed. They have very bad accuracy. Full speed to get close to the part then VERY slow to get closer. Our hydro vise opens about .100" and we need every bit of it of that tells you something. BUT with all of that being said, with high quantity jobs it for sure pays. For an example we just ran a job that would take 3 hours to setup and about 60 hours of man hours in front of a machine. Well I set it up with the robot and have 6 hours of loading the "matrix"(table with parts that the robot picks from) and the job is done. Cost per hour on that mill is less than half of what the others are. So short version, it pays off of you have the proper setup and quantities to do it right. Hope this helps!

Hi, this does help but comment and a question. First, just FYI, Baxter was the predecessor to Sawyer. Question: While it's great you found an opportunity to save money, is the cobot truly performing an operation that it is the most cost-effective for? To answer this, are you 100% certain that it was the "cheapest, easiest and most productive" motion available? As a possible alternative, what about a few linear actuators and a PLC (depends on a lot of factors)? Okay, have to do some design and integration but, once that's all done, could be WAY faster & accurate, cheaper and much easier to set up if this is a job you run on occasion. Here's a related question: did you end up using it just so you could say you put it to use? I didn't fully get all of the "politics" involved in your situation. I've had projects in the past where I worked for a German-owned company (in the US) and we had to use certain equipment that was clearly not the "most suitable" for the application so I totally get that stuff like this happens.

Thanks and I hope you don't take this as criticism, I'm really just trying to "get" the scenario.
The Dude

So far that's been my conclusion every time I've started to design a cobot into a work cell. It's always been cheaper, faster, and with better accuracy to use either linear actuators or a traditional SCARA or 6-axis robot, even with the cost of guarding. I haven't given up on cobots and still look at them in every scenario but they just haven't made sense for any of my applications yet.

Originally Posted by Punkinhead

So far that's been my conclusion every time I've started to design a cobot into a work cell. It's always been cheaper, faster, and with better accuracy to use either linear actuators or a traditional SCARA or 6-axis robot, even with the cost of guarding. I haven't given up on cobots and still look at them in every scenario but they just haven't made sense for any of my applications yet.

And from the "opposite" end, I think a lot of people involved in cobots (especially on the sales end) don't know about other options and so that's the only thing they know how to sell. I recently attended a gig where the main speaker clearly only knew primarily about cobots (not even standard industrial robots). His things was "walk out into your factory and look for easy applications". If there's a factory with engineers who can learn about cobots and then spot a lot of easy applications then they have been missing out on labor/money savings for years. All of the easy stuff should have been done a long time ago, with or without cobots. My belief is that it's automating (or semi-automating) the more difficult processes that gets you ahead of the competion...stuff other people would scratch their head over. Not saying it shouldn't have a favorable payback, just might not be a "super-simple" solution.

Tu summarize the low automation skills for coboters, many appear to be millennials who think that cobots are "cool" and are going to take over the automation world and haven't taken the time to investigate alternatives. I spoken with a few sales people at distributors for cobots and they seem to fit that category.

The Dude

I would love to hear stories about cobots too. I haven't tried it yet but I hope I'll be able to experience its revolutionary efficiency in the industry. It's quite new and expensive for medium business.

Originally Posted by julkor

I would love to hear stories about cobots too. I haven't tried it yet but I hope I'll be able to experience its revolutionary efficiency in the industry. It's quite new and expensive for medium business.

Thanks for your interest, but one caveat: I would have to argue against "revolutionary efficiency" with the sole exception being that they are easy to program and (therefore to) re-deploy. I would say that, thus far, there is a very narrow range of applications for which they are "efficient and effective" for. The one exception to that is for load/unload times with the following requirements: It either has to be a very long cycle time where the "interruption" of the load/unload is negligible or it has to be loaded/unloaded on a change-out basis (the cobot can load/unload on a station while the machine is running another part. The problem with this latter strategy is that, if you have to make a custom mechanism to allow the cobot to do this, then you might be more cost-effective in making a custom "complete" load and unload device. Hope that makes sense.

Thanks,
The Dude

The Dude - I think what you are saying is that, so far, cobots are not showing anything like the speed and precision of standard robots, gantries, etc. So they're only a win in configurations that change often and where an unattended slow load/unload is OK. (Or did I misunderstand you?)

Originally Posted by bryan_machine

The Dude - I think what you are saying is that, so far, cobots are not showing anything like the speed and precision of standard robots, gantries, etc. So they're only a win in configurations that change often and where an unattended slow load/unload is OK. (Or did I misunderstand you?)

Yes, but I need to clarify that's MY understanding and it is with somewhat limited experience (with cobots, not automation in general, which is a big reason why I made the post). The other caveat is that there are brands of cobots out there that are clearly "better" (precision, load and possibly speed) than others (of course, likely more expensive). I'm going to Fabtech this year and will definitely be studying whats out there.

Bottom line: any "investment" should have alternatives (and not just "doing nothing"). For a cobot, weigh it against making a custom mechanism. The benefits of a cobot cell would likely be "easy to purchase and program", "re-deployable", etc. but it might be more expensive and less capable than a custom device.


The Dude

Thanks for all your help guys. I went with the Lenco 02080, it was only about $130 on eBay and basically a bolt-on part, very easy to integrate. Only downside it's rated for up to 500 amps so it had a 3/4" thread but was able to accommodate with a larger shaft.

The Dude