First robot suggestions?

[Automaton]

Hi,

I would like to get my feet wet with a robotic arm. My current motivation is to run post-machining processes on turned parts coming off a small lathe.

I've been looking for a decent starter machine. Ideally one that is small, not too expensive, and which uses an interface that will be transferable to larger more powerful machines in the future.

It seems like the Kuka/Fanuc/Abb lines don't really extend down into a "desktop" size, which would be ideal for me. I would prefer to have this thing occupy 16 ft^2 not 200 ft^2

Lynxmotion appears to make small robot arms, but their marketing feels much more hobby/toy/educational than real work... which is what I want to do with the thing.

Perhaps the MG400 is about right? DOBOT MG400 Industrial Desktop Robot Does any one have experaince with these?

Then there are a bunch of open-sourced/3d printed things. Everything I've seen like this feels crappy, but at least I could redesign & print new parts as they fail.

Bill

 

[jtaylor]

I'd look hard at the UR-5 from Universal: It seems to be about the cheapest real robot (still its around $30K last time we bought one). I've used several and the quality and user interface are very decent.

It takes a lot less time to train them than other robots so it cuts down on frustration from normal users.

Collaborative robotic automation | Cobots from Universal Robots

 

[CarbideBob]

First robot... fun stuff.

 

[DavidScott]

Epson seems to make the best for what you are looking for. Here is a link I saved on it, Epson's website has a lot more info on them. The last I checked it was under $16k. I have heard way too many negative stories about Universals and the joints are not sealed so you can't use them around coolant. Universals also don't have any internal air or electrical circuits for end effectors.

 

[Strostkovy]

OTC has some neat robots. I haven't looked into their little ones but the welding ones we demoed and purchased (but haven't received yet) are pretty great.

Here is the smallest robot I see: FD-H5 Compact Robot | OTC DAIHEN

As I would expect with any robot, your best bet is making your own custom end effector.

 

[jbs007]

Our first was a Fanuc M16. I know it sounds bigger than you want, but realize that the whole work envelope is not as useable as you may think, not currently having a robot. when that arm starts extending pretty straight, the joints are moving quick and it doesn't seem good for the robot or the job. The robot interfering with itself (at all the joints) can be an issue when you are working too close to the base, then you also have singularity points you need to avoid, etc. I am not a robot pro, but what I know now, I would not expect to comfortably use the advertised envelope of a robot.

We are probably hard on our robot and need it to traverse a long distance accurately, which may not apply to you, but when I can I like to avoid work near the outer edges of the envelope, work to close to the base, and work straight in front. Ideally you may consider a robot where 90% of your work can be done well within one quadrant of it's envelope.

 

[Strostkovy]

Look at the range with posture, which is basically where the robot can reach with the end effector straight down. Full extension is unpleasant to program at.

I think the term mostly applies to welding robots, but should be somewhat universal.

 

[michiganbuck]

First robot suggestions?

Be sure that your Robot knows that you are the Boss.

 

[matt_isserstedt]

There are a lot of Fanucs coming out of the auto industry. M10iA is a nice small place to start. Everything you learn will transfer up to larger sized Fanucs.

Good to learn about the tool/user frames...those are the most powerful features imo.

Then you need something to interface with the robot...maybe its also time to get a PLC and some ladder logic programming. Granted the robot can do simple tasks on its own, but the I/O and programming is not nearly as fast to my mind.

Critical in Fanuc that the robot arm and robot controller have the same "F number" which is a serial number. Don't buy one without the other.

 

[Strostkovy]

Our OTC came in two weeks ago and we started welding parts with it immediately. I think it's nice. Controls are outdated, but that's the case of every robot I've seen.

It cost $80k with about 5' usable reach, a 500kg positioner, and a welder and all required hardware, so the smaller units may be pretty affordable.

I am not familiar with their material handling line but it's what I'll be looking into when we increase our automation.

 

[SVFeingold]

How much reach and payload do you need? Don't forget the end effector.

All of those brands you listed make small desktop-sized robots. There's also Universal Robots, Staubli, Epson, Omron, Yaskawa, Rethink, Kinova, the list goes on. Brother even makes one now. For many of them the controllers are far larger/heavier than the robot arms themselves.

As far as open-sourced/3D printed, eh...most are your standard maker-tier garbage. More for fun and software development. If you want a fun project and don't mind the extra work in slowly replacing broken parts with better parts then that's one way to go, but make sure that a project is what you want in that case. Speaking of software it will be much nicer to use/program a robot that comes with a real controller and software package which is built-for-purpose and easy to train. Open-source stuff is a wildcard. 300 people working mostly independently with no clear direction or leadership do not generally churn out consistent and well-integrated products. It depends on your budget and what you want to get out of this project and/or where you'd most like to spend your time.

 

[e30ryan]

I have been working/programming an ABB robot for the past few years. This is the first robot I have ever touched, but I have been fairly pleased with it. The included RobotStudio software is easy to use and the programming language is also quite easy to learn (similar to C languages).
We bought a relatively large robot (IRB6700), but the price (~$80k USD?) was very reasonable for the capacity/capability you get with it.

 

[michiganbuck]

The scientists behind the brain-computer interface technology are now seeking funding to provide similar implants for other people with ALS, which will cost close to $500,000 over the first two years of use.

Likely we will soon be able to make a small organic computer out of a brain, put it in an autonomous truck and lay off all the truck drivers.

Walking robots powered with these brain computers could do all the jobs, and then just be unplugged after each shift. That would really lower the cost of products, and eliminating organic people would lower the need for most products..
So at some point in time, we could just unplug all the robots and save all the energy. likely get the last robot to do that.

Likely use babby brains so they are not cluttered with a lot of unnecessary thinking.

 

[mrclauds]

We have 4 fanuc robots each with a 2 meter reach envelope. Each robot feeds and unloads 3 cnc lathes. We have had 1 break down since 2008, as they are incredibly well built.

However, we have built enough of our own custom machines and electronics to know that we can build a robot for a tiny fraction of the cost. That being said, the forward and inverse kinematics for calculating the movement and trajectory of 6 revolute axis's is a nightmare (at least for me, and a main reason why hobby revolute robots will not do decent industrial work). So we are building linear robots instead.

As for the maker style electronics, the main difference between a good linear robot and a bad linear robot is the frame and quality of linear movement devices(linears, ball bearing blocks, servo motors etc).

Look into linear robots, easier to make than you think, then if you run fanuc controls on machines with macro b, you can use the control itself to control the robot instead of having a seperate controller

 

[Wick Craft]

I thought about dicking around with one of these to feed a VF-2 with Air Vises: ROBOT KITS | Annin Robotics

Can be fabbed with a 7th axis track to drive into the enclosure. Dude sells an aluminum frame kit for a reasonable price. Looks like it can be "all in" for $3-5K . I believe it can do a ~4 lb pickup with a 24" reach.

Vids:

AR4 6 Axis Robot Arm - YouTube

3D printed version feeding a little Tormach

AR2 / AR3 robot CNC tending - YouTube