What's new
What's new

What is involved in adding a robot to an old VMC? Keep it as simple as possible!!!

DavidScott

Diamond
Joined
Jul 11, 2012
Location
Washington
I have a 1996 Enshu S300 drill/tap mill that I don't use much now, no 4th which is very limiting for me. It is quite old and fairly loose but still holds tenths, is dead reliable, and fully supported by the MTB. I "registered" it so I have unlimited phone support and 3 full sets of all documentation. I have about 3-4 months a year of 2-1/2 axis work I could put on this mill if I could add a robot to it. They are all aluminum widgets in 5 sizes but out of 3 bar sizes, 1/4"x1" 4" and 6-1/4" long, 1/2"x2" 6" and 8" long, and 1/2"x3" 8" long, all real simple parts machined in 2 ops. The mill had a rotating table option, not on mine, so I have a great mounting surface with threaded bolt holes to attach the robot mount to. Anyway, I have been working towards a robotically loaded VMC for a few years now and think this may be the start. I have no experience with robots and figure I would start here with my questions, I also figure I am not the only one here wondering about this. I really am most interested in the robot to old VMC integration right now. I would like to save the automatic vises and EOT for later, not that I haven't been thinking of it. The mill is in one end of my shop and there is only one path to access it that I could control with a light curtain, or such. Please keep in mind I am a one-man shop working from home with your solutions/suggestions. I would like to keep this as simple and cheap as possible. Budget depends on what is needed. If I need to buy new to get support then so be it. I am thinking of a 1ish meter reach 5-7 kg payload robot. My heaviest part weighs 1.2 lbs, so 3 weigh 3.6 lbs, which it will have to hold at one time to swap out op1 and op2 parts in one cycle.

Time for lunch so let's start here. Yeah, I figure this thread could cover a lot of territory. No secrets about the parts so I can fully document any of this process if I end up doing it.

Initial questions are:
1) Robot to VMC integration, what is involved?
2) Where and what to buy robot wise? New? Used? I remember, I think, Tony said they like Kawasaki best.
3) How much support from the robot MTB? Whether new or used?
 
Last edited:
I've done a pretty budget integration with a Haas and an ABB robot. I'm not the world's foremost expert, but it IS running parts right now while I'm in my chair at home.

At minimum, you'll need a signal from VMC to robot, a signal from robot to VMC, and appropriate safeties.

Haas has M code fired relays. We chose one and this is our signal to robot. M51 tells robot to load.

Found the M-FIN input on the Haas board and this is the signal from robot back to mill. After we command the M51 the machine waits until the robot sends back the M-FIN signal when the robot has finished loading.

We bought a used ABB IRB 1400 off of Ebay. It has run like a champ and it's a 1998 model. Even though they've never made a dime from me, ABB has supported me like I own the only robot in the US. They've been incredible and I'd have a hard time not buying ABB again.
 
That sounds simple enough. I was thinking of tieing into the end of cycle signal and cycle start button as the most basic until I try using proximity sensors in the vises. Glad to hear about ABBs support, we all know how important it is!
 
Abb is great for support same with nachi.

It's a piece of cake to install a robot to a machine tool.

Sent from my Pixel 3 using Tapatalk
 
For small budget projects like yours I don't see any reason to buy used. Epson has their VT6L model that is only $14k. 6kg payload, 900mm reach, and the programming software is included with the robot. Manuals are freely available on their website as well, unlike many other robot manufacturers.

Fanuc control on the Enshu, correct? Do you have any free relays on the machine? If not, you might need to have the ladder modified and/or install an additional IO module.

The vises can be done completely through the robot and the machine would be blind to them. I would just add a button on the operator's panel to open/close them by hand.
 
How robust is the Epson? What does it give up over the more expensive robots? In many ways it does look like a perfect fit.

It has a Yasnac J50, their version of the O control. I don't know about extra relays, it is a pretty basic production drill/tap with no extra frills.

Oh, and do you have any affiliation with Epson?
 
How robust is the Epson?

It's a new model, so it's a bit early to tell how well it'll hold up.

What does it give up over the more expensive robots? In many ways it does look like a perfect fit.

Speed and options. There are a few videos of it on youtube moving and it's not a snail, but it won't be breaking any speed records, either. If you're worried about the robot cycle being 30 seconds vs 15 seconds, it isn't the right application for it.

You can get vision for it, or expand the IO via EIP/PB/EtherCAT, etc but they've basically taken the Haas model of giving you the base unit and options are extra. If your curious about the costs my local distributor in Wisconsin, Doig Corp, has prices publicly listed.

Oh, and do you have any affiliation with Epson?

Nope. If you could say I have any affiliation with a robot manufacturer, it'd be big yellow. I'm deep in their ecosystem, but I would recommend you stay away from them due to your lack of experience with robots and Fanuc's secretive bullshit re used robots.
 
Just buy a nachi and never look back...

I have each and fanucmates now and the nachi robots are most accurate, fast as lightning, and very reliable.



Sent from my Pixel 3 using Tapatalk
 
I'm working on similar right now. I recently bought a Haas VF-0 for $1 and instead of flipping it I'm going the robot route. I bought an Intel silicon wafer loading robot for $30, but no software to run its obsolete control. A friend has developed an automation HMI and is currently repowering the wafer robot with an Arduino brain.

The goal is to prove out his automation stuff, run all my products and his products that will fit in the VF-0 envelope.

If the wafer arm doesn't work out (it has reasonable build quality, but uses steppers) I will buy whatever properly sized used industrial arm is cheap on Ebay.

Some of my parts are pretty heavy. 10+ lb blanks.
 
...
It's a piece of cake to install a robot to a machine tool.
Interfacing a robot to unload/load is a piece of cake. Cycle stop, grab, get new part hit the green button. You can literally do this with a photocell and a tiny air cylinder and the machine tool will think a human is out there.
Making the robot do it's part of the job of place output and grab the next new part is not so easy.
For more fun do bin picking on the going in side.
For some the problem becomes fixturing the input and output side so that the stupid robot can just pick and place.
That part can get expensive even if working out of standard dunnage if more than one part family.

Safety rules.... OMG, I think most small operations could never do it all.
Do what you can. Light curtains, cage switch and pressure mats telling the robot no a good start.
Light curtains expensive, cage and reset or the lockout a pain and more money.
One man shop and you can get away murder on all this safety stuff.
Bob
 
Interfacing a robot to unload/load is a piece of cake. Cycle stop, grab, get new part hit the green button. You can literally do this with a photocell and a tiny air cylinder and the machine tool will think a human is out there.
Making the robot do it's part of the job of place output and grab the next new part is not so easy.
For more fun do bin picking on the going in side.
For some the problem becomes fixturing the input and output side so that the stupid robot can just pick and place.
That part can get expensive even if working out of standard dunnage if more than one part family.

Safety rules.... OMG, I think most small operations could never do it all.
Do what you can. Light curtains, cage switch and pressure mats telling the robot no a good start.
Light curtains expensive, cage and reset or the lockout a pain and more money.
One man shop and you can get away murder on all this safety stuff.
Bob
Installing is the only easy part lol. Programming and every other aspect is a pain in the butt... Still less of a pain than employees.

Sent from my Pixel 3 using Tapatalk
 








 
Back
Top