What's new
What's new

Any good/bad cobot stories?

The Dude

Hot Rolled
Joined
Oct 19, 2010
Location
Portland, OR
I'm going to be involved in a process that evaluates potential automation opportunities for collaborative robots. Two points I should make clear up front: First, I have years of automation experience going back to robotic welding in the late 80's. This includes "standard cells", completely custom systems, robotic, non-robotic, etc. Second, I made it clear to the organization I'll be working with that I'd be more of a "filter" than a "pusher". I mostly stating this as I don't want anyone to think I'm trying to get free sales data or things like that. Short story is I'll be helping companies make "good and smart" decisions and be on their side, not the integrator/selling side. The main organization involved is a personnel/staffing company. I was a bit shocked when I heard this but then realized this is a "let's join rather than fight strategy" which I understand.

My overall view on cobots, and a lot of other new technology (Industry 4.0, 3-D printing, IoT, Six-Sigma) is that it's "solutions looking for problems". All of these have valid applications but a lot of the time the sales people really push hard and you might overlook a more suitable traditional technology. For example, 3-D printing has clearly shown benefits for packaging material, spare parts, medical, prototyping, etc. but hasn't become quite the game-changer that some people thought it would be (at least in my view). And don't get me started on Six-Sigma...(great way for consultants to make money, IMO).

What I'm looking for are really short stories about any experiences with cobots. For an actual implementation, it would be great to hear what was the process, what were the savings (labor, quality, safety, etc.), and who did the integration. Finally, in the end, was it successful and/or would you do it the same way again? You could say all that in just a few sentences. If you don't mind, I'd also like to know the brand of cobot. You don't need to provide any details or specifics about your company.

For investigations that didn't proceed, I'd like to know what the issues were: cost, lack of capabilities, unstable process, too scary, salesperson was a jerk, etc. If you could mention something about the process that would be great.

I am reasonably familiar with the whole cobot industry. I saw early demos on Baxter, worked for a supplier of parts to that industry and have been watching the industry a bit (IMTS, Fabtech). I do expect over the next few months to get reasonably up to speed on both the supply and demand end. Feel free to PM me if you have any questions and I'll do my best to give back on this effort.

Thanks for any input.
The Dude
 
"For investigations that didn't proceed": here is some experience with cobot sales associates, I contacted the "first" sales guy who is employed directly with Universal Robot the manufacture, I asked for a demo, he shows up with a non-working unit, set's it up and it will not even move, a waste of 4 key people's time, the WTH here is he knew it was not working when he drove 4 hours to our location.
"Second" sales associate works for the area distributor, he emails me wants to come to do a demo, I replied, no thank you, I do not need to see the robot operate. He replies he is going to demo a touch feature plus walk us through the easy set up of a certain part we are buying the UR10E for.
So OK, he shows up, set up and after about 30 minutes of fumbling around, he tells us he has not been trained on that cobot.
As he is packing up, I tell him to give me a price on an air spindle attachment, it's been days now and no follow up email, message, quote....nothing.
We have gone from ready to write the check for the unit plus end of arm attachments to, rethinking this purchase.
 
Hey, totally not surprised by that (unfortunately!) and I do feel your pain. Definitely a few factors at play here. I think the sales end of cobots is way ahead of the "reality of application". Aside from simple & repetitive part loading/unloading, maybe packaging, I'm just not seeing much opportunity. I'm still getting up to speed on the latest & greatest but when I attended a "Baxter" demo (Rethink's first product) about 5 years ago, it was pretty laughable. No wonder they went out of business, they're sales were mostly to academia. Wonder the German company that bought them will do much better with this product line.

There is also a whole new group of people involved in the sale of new technology and I think it's less "robust" than traditional technologies (like industrial robots, PLC's, new machinery, etc.). Yes, lots of "shiny-shoes" in that arena as well but generally more focused on "providing a solution to a problem" that "a solution looking for a problem" like new tech is.

Thanks for the post. A bit surprised I haven't had more responses, I really tried to word it in a manner to that it was short enough but also enough detail so that people would know I wasn't trying to sell them something or "whatever". Just looking for stories, that's all.

Thanks again,
The Dude
 
Thanks for the post. A bit surprised I haven't had more responses, I really tried to word it in a manner to that it was short enough but also enough detail so that people would know I wasn't trying to sell them something or "whatever". Just looking for stories, that's all.

There have been quite a few people talking about their experiences in various past posts. Your best bet is probably to look at those. All UR robots, IIRC. I don't think I've seen any hands-on stories about the FANUCs, Kukas, etc.
 
There have been quite a few people talking about their experiences in various past posts. Your best bet is probably to look at those. All UR robots, IIRC. I don't think I've seen any hands-on stories about the FANUCs, Kukas, etc.

I should have mentioned that I did look at prior posts, they seemed to mostly be about load/unload. I believe I meant to clarify that I'm especially interested in anything beyond packaging/loading, e.g. assembly or speciatly processes. I'll go back and edit that.

Ooops, won't let me edit it anymore. But yeah, especially interested in stuff beyond simple load/unload & packaging!

Thanks,
The Dude
 
horizontal cnc's that share 24 pallets (44"square tables) that a robot is used to put in racks with parts (parts often weigh over 1/2 ton) in fixtures on pallets or load pallet on to cnc or bring to setup station
.
the pallet mover robot is heavily dependent on people, it tends to get stuck or experience malfunctions. originally designed for lights out machining. what types of malfunctions ? many but mostly a loosing home position (off a 1/3") and not registering safety locking pin engaged so robot can move away from unloaded pallet in storage racks. maybe 10% of time got to walk over and wiggle pallet and sometimes have to use prybar cause robot not aligned close enough to unload to rack.
.
i am told when new it worked ok but over the decades it experiences more and more malfunctions. when anybody gives you a 3 year or even 10 year warranty it can go by extremely fast.
 
horizontal cnc's that share 24 pallets (44"square tables) that a robot is used to put in racks with parts (parts often weigh over 1/2 ton) in fixtures on pallets or load pallet on to cnc or bring to setup station
.
the pallet mover robot is heavily dependent on people, it tends to get stuck or experience malfunctions. originally designed for lights out machining. what types of malfunctions ? many but mostly a loosing home position (off a 1/3") and not registering safety locking pin engaged so robot can move away from unloaded pallet in storage racks. maybe 10% of time got to walk over and wiggle pallet and sometimes have to use prybar cause robot not aligned close enough to unload to rack.
.
i am told when new it worked ok but over the decades it experiences more and more malfunctions. when anybody gives you a 3 year or even 10 year warranty it can go by extremely fast.

Just to clarify, this is likely not a cobot but a standard industrial robot (as you mentioned). One thing I have noticed is that product support for robots (from the manufacturer) tends to be on a shorter life cycle. Two reasons for this: 1) they want to sell you a new robot eventually and 2) I suspect the vast majority of robots (at least the largest customers) are in the automotive industry (or similar high-volumes) where product life cycles tend to be short. After 3-5 years, they are replaced due to an entirely new product line.

You mentioned "over the decades"....sounds like it's time for replacement. Chances are that it's no longer supported by the manufacturer.

The Dude
 
many a machine especially robots works ok when new. when older and dirty and worn out you get the errors like getting stuck, out of position, etc
.
all good to say replace robot. some robotic systems are easily over $1,000,000. when large parts need replacement its not easy to get parts. whats worse is electronic parts no longer made. circuit boards that malfunction for many reasons i have seen on many old machines. talk about not doing what it is suppose to do. even sensors reading electronic chips if the chips were only $100. few order a extra 1000 chips for later replacement.
.
electronics parts not available i would say is worse problem than mechanical parts. mechanical parts often you can get a drawing and remake part. electronic chips good luck getting a drawing on it to make custom electronic parts. and being told we dont give out drawing cause it is propriety or secret design is common.
.
i even asked a rubber manufacturer once can i get a list of ingredients in the rubber to see if any would react to solvents i was using only to be told its a secret recipe they dont give list of ingredients. many manufacturers do not give out drawings even when they no longer make the parts.
.
you buy a car do you think car manufacturer going to give you a complete set of drawings to remake every part in the car ?? a perfect example you have a iphone and you want to make parts to fix a mobile or cellular phone that is not working...... good luck with that. short of setting up a billion dollar factory and hiring thousands of experts to remake electronic parts...... i have heard it called the impossible project
.
recreating what was done decades earlier without detailed procedures cause people didnt write everything down is not easy. many cannot even recreate parts they made themselves a decade earlier cause they didnt write stuff down and forgot what they did. i have seen that often enough
 
i get a laugh when people think it is easy to fix any machine. take even a mechanical Rolex watch, a real one. say watch stops working and you need to repair and or replace parts. good luck with that. sometimes the smallest parts take the longest to make
.
even robots can have highly complex custom parts. just like a new car you have few problems when new but when car is 10 to 20 years old you get random failures of different parts. old robots are similar. not easy to fix or maintain. and very few buy a lot of spare parts same as you dont buy $10,000 of spare parts for your car figuring you will need them 10 years later
 
I am reasonably familiar with the whole cobot industry. I saw early demos on Baxter, worked for a supplier of parts to that industry and have been watching the industry a bit (IMTS, Fabtech). I do expect over the next few months to get reasonably up to speed on both the supply and demand end. Feel free to PM me if you have any questions and I'll do my best to give back on this effort.

The programming on the UR robots is excellent.

That said, the original Baxter had two arms. I'd like to see a 2-arm UR connected to the same control box. I'm not all that interested in syncing two independent URs.

I'd also like to see a UR20, i.e. 20kg payload. If two arms can sync to get to 30-40kg payload, even better.

Lastly, I'd like to see automatic end effector changes, even if it means adding an extra component.
 
some robots if not lined up get stuck cause a locking pin wont go in confirming position. if locking pin is tapered or bullet head shape it can go in hole and push the parts into alignment.
.
having a old robot stuck and wiggling it or hitting it so parts line up gets old after awhile. doing it 10 times per day is flat out annoying. some design changes can really help with reliability as machines get older. when the Fonze on Happy Days TV show hit the juke box to get it working people laugh. it aint funny when you have to do it yourself every day many times per day especially ff you have to climb a 20 foot ladder carrying a pry bar and crawl over oily racks not meant for walking across
 
I'd also like to see a UR20, i.e. 20kg payload. If two arms can sync to get to 30-40kg payload, even better.

Usually that kind of thing is an automation no-no (my work has never been on 6+ DoF robot arms though). If they can get their software really tight it might be an option, but it's usually better to just build a bigger machine.
 
Usually that kind of thing is an automation no-no (my work has never been on 6+ DoF robot arms though). If they can get their software really tight it might be an option, but it's usually better to just build a bigger machine.

Sure, but most of the time we're talking two independent robots. Here I'm referring to two arms connected to one controller, thus no syncing/runaway issues. As far as the controller is concerned, 12 axes on one arm, or 6 axes each on two arms, should be the same.
 
The programming on the UR robots is excellent.

That said, the original Baxter had two arms. I'd like to see a 2-arm UR connected to the same control box. I'm not all that interested in syncing two independent URs.

I'd also like to see a UR20, i.e. 20kg payload. If two arms can sync to get to 30-40kg payload, even better.

Lastly, I'd like to see automatic end effector changes, even if it means adding an extra component.

Look here- End-of-Arm Tooling for Cobots | New Scale Robotics Down the page a bit, we can mount you up a triple set!
 
Collaborative robot.

Sure, but most of the time we're talking two independent robots. Here I'm referring to two arms connected to one controller, thus no syncing/runaway issues. As far as the controller is concerned, 12 axes on one arm, or 6 axes each on two arms, should be the same.

Syncing is one issue, but mechanical failure is the other. You have two arms supporting something, if one is damaged already or breaks under load, the load then breaks the other one, and the load goes who knows where.
 
Ok I have to ask, what is a cobot?

It's also short for "collaborative robot". While I'm not familiar with all of them, the ones I am are quite slow (and low of weight capacity) which I believe makes them less collaborative than they ideally should be. If it moves significantly slower than a human, it limits the type of operations where it can work in that fashion next to a human. Even for part loading/unloading, if it's slow then it may slow the machine down as well, unless it can do the load/unload at an idle station of a rotatory table or something similar.

The Dude
 
Ok I have to ask, what is a cobot?
I think the big difference is cobots have load sensors on each axis. If working in cobot mode they are 1/2 speed or so and if a sensor detects a larger than anticipated load it will stop. The load sensors are handy in "pressing" the part to the stops before clamping. Granted I have never worked with any robot but am paying attention, as I would love to get one to put me out of the job of production operator.
 








 
Back
Top