What's new
What's new

automation design question - use for product

Jason H

Stainless
Joined
Mar 29, 2006
Location
Los Angeles, CA.
I am in love with this new forum. Thanks to all who contribute here.

I have a widget that I need to automate for sewing. It is a piece of fabric that is extremely stiff, almost like cardboard. It is cut into many different shapes based upon customers need. This is done on a die cutter ( think giant cookie cutter ). The maximum size is 18" X 18". Widget weight is under 1 pound each. the cutting part will be done by hand.

From there it is sewn around the edge. The sewing time is extremely long based upon the cost of the unit. I want to automate the sewing part, here is what I have thought of:

From Die cutting it is placed into a stack. It is picked up through suction, and moved over to the sewing machine. PLC is timed to lift sewing foot and widget is placed underneath. The arm needs to move at about 1 inch every ten seconds ( based upon sewing speed ). There are some starts and stops as it turns corners or makes a radius. The arm will need to maintain a perfect plane so that it is sewn straight. Sewing cycle is over and the widget is moved from under sewing foot to drop box.

Here are my questions. Please also note budget is an issue. My Factory is outfitted with 440V @ 1200 amps. My machine shop is for MRO and R&D. I have two Monarch 10EE's, two Tree 2uvrc's, and a clausing 5914. All parts will be made by hand.

For some of the other operations, I will want to use a PLC, can a robot arm integrate with a PLC?

What arm would be suitable for this application? Size is not important, this will be all caged off once operational.

I will have a motor directly driving the sewing machine. It will have to start, quickly stop, and change speed. It will also have to know position to help translate if the needle is up or down. What type of motor is best for this?

Where can I see the different types of attachments for robot arms?

Can the arm path be controlled by an autocad layout?

Please treat me like a two year old, I am just starting out with this stuff.

Jason
 
I have automated several sewing machines for production sewing -

1.) We used standard 3/4HP Baldor Induction motors for the sewing machine operating in closed loop vector mode and a small prox switch on the machine (upper part of reciprocating shaft that holds the needle) for homing / stopping with the needle in the up position.

2.) You will need a quality tension controller for the thread / string you will be using

3.) You will need a thread break detector so you dont produce a bunch of scrap

For the robot - a simple smallish SCARA robot with a rotary end-effector should be fine for most any geometry.

Since you want to coordinate the Robot velocity along a path with the speed (or more correctly the stitch count) of the sewing machine - the encoder on the sewing machine motor should act as a feedrate source for the robot motion controller. Program the robot to advance a certain distance per stitch and then run the sewing machine at whatever speed works best for the process.

Your challenge is going to be finding a robot motion controller that will manage all of this seamlessly (pun intended) . . . LOL

If it were me - I'd buy a robot of suitable geometry and payload capacity and put my own controller on it - for something like this, I think you will need a 4-Axis SCARA (3-Axis on the arm + elevation) and a 5th axis to control the sewing machine.

Engineering / Programming / Hardware for something like this off the top of my head will be on the order of $75k - $100k unless you want to source stuff from eBay - better figure 160 - 240 hours of engineering alone.

If you are die cutting - then presumably you have CAD data on the die which can then be used to create a motion profile for the robot. It is fairly easy to design the SCARA robot to keep the perimeter of the part feeding toward the needle - and the rotary actuator on the end of the arm that is holding the part can be programmed to keep the part feeding tangent to the path.

The sewing machine over-arm will need to have a deep enough throat to accommodate the largest part swinging about any point along the perimeter - if this isn't possible, the programming gets a lot more complicated to anticipate when to stop / disengage - rotate - re-engage, etc.
 
Thank you for the response. I have cut and extended the machine so space under the head is not an issue.

For the thread break detectors, I was hoping for something optical versus the old school mechanical. Do you have any suggestions?

I believe that I would also like to be able to do larger sizes, on the magnitude of 90" X 90". With the need of a little bit more complexity. Some of the decommissioned larger arms seem to go pretty cheap. Any reason not to use them?

Great idea on changing the motor speed on the sewing machine to match the robot. If this is all automated, I can run this all day long and not worry if the sewing machine is running under capacity. A VFD with dynamic breaking would work for this to get speed correct and provide a near instant stop.

For the tension controller, I thought I could use the rotating type that are on embroidery machines. There is a thread break sensor on there that can also be tied in. The optical sensor would also work if the thread ran out. Do you know of a non traditional style thread tension device?

Once the speed of the robot is matched by everything else, it seems as if this would just be a timing issue for each step.

Thanks for the help!

Jason
 
We already do this and for a LOT less than $100k. It's not terribly complicated to control most any device from a standard robot controller. I'm a big fan of Fanuc, and Model B I/O whcih makes hooking up the I/O a breeze. There are loads of used 6 axis robots out there for under $15 k that will do the job.

You can mate them to an industrial sewing machine that has automatic neddle positioning, lockstitch and backtack, and time the robot motion through the normal shaft encoder, or you can make life simpler.

Use a programable pattern tacker to sew out the entire pattern (Think Used like Brother 311 or 432 depending on whether you're and X/Y or Y Theta kind of guy). Your robot then needs to simply pick up the part, drop it off, hit cycle start, and wait for it to finish before unloading the finished part and repeating the cycle.

You can easily incorporate thread break, but it's pretty rare with a pattern tacker. You will still need to figure out a way to change out the bobbin. I've never figured out how to do that robotically yet.

If you're any kind of a hack at all you can do it for under $20k easy.

Stu
 
Oops, skip the pattern tacker, I just noticed the 18" x 18" sewing field.

That makes life a bit tougher. Your going to need to interface with an embroidery machine, or move the fabric with the robot....
 
Okay there are two more sewing people on this board than I thought there were. I have several of the JUKI AMS series. I bought the widest field that Juki had from a parachute mfg. This is still smaller than what I need.

I have a mentor that is extremley brillant in both sewing machines and machining. With that part covered, I am the one doing the programming which I dont have a problem with.

there are several pattents that show bobin changers, but I am finding that my mind gets poisoned on design when I see something else. Might be best not to look. Some of them are super old so you might be able to improve on them.

What is best to look at when inspecting a used robot? Where are the wear points?

I have been going through HGR, Ebay, and Automation Direct looking at different things. Any idea on other places to shop? I also have a motion industries location within walking distance from my factory.

Thanks again!

Jason
 
Also for the single needle operations, do any of the manufacuters, such as Juki or Brother have an interface that a robot could tie in to on their sewing machines?
 
Jason

We tried an optical solution, but it required re-learning each type of thread due to different optical characteristics - we used a sensor from Keyence and ended up going back to mechanical with a prox switch and a debounce timer in the controller.

- also, what kind of safety system does your company typically require? Do you have any experience with OSHA audits with robots / automation equipment?

Our experience has been that on something like this you will need perimeter guarding with the infeed / outfeed staging areas behind some kind of laser curtain or pressure sensitive mat. If you don't have these requirements - you need to make some effort (and document your methodology) to ensure that an operator cannot be injured by the robot while feeding the process with blanks or taking finished goods away from the work cell.

If the parts you are sewing are roughly 8 feet in diameter - how stiff / self supporting are they?

Stuart - would you have any interest contracting Fanuc Robot programming locally? We have a customer with a new Fanuc that he is not at all happy with - we have been engaged to redesign the end-effector and I am seeing some significant problems with the way it was programmed by the OEM - (picks up pallets, slip sheets, and stacks product for shipping).

As for cost - I couldn't touch it for $20k - schematic development, UL508 compliance component listing/audit trail, programming w/documentation, manuals . . . that would take more than $20k without buying a single piece of hardware. :)
 
Thank you for the tip on the optical sensor.

for saftey factors, I was going to cage off a good part of the unit and light curtain in the other area.

For the smaller parts that was orginally thought of for this project, they are extremley stiff, but the larger parts are single pieces of material that are pretty limp. I was thinking some sort of vacuum plate to hold them. Any better idea?

Thanks

Jason
 
It appears that Banner Engineering is still in business:

http://www.bannerengineering.com/en-US/products

I use some of their photo-optics for part detection in plastic injection molds (to ensure all the parts are ejected before the mold closes again) and their products, while not cheap, are reliable and bullet-proof. I seem to recall thread detection was one of the uses listed for some of the smaller photo sensors, but its been a long time... Still worth checking out, I think.

Dennis
 
Dennis, I have seen some of their thread sensors used on ebay, but could not find any info on the acutal sensors listed. I have to imagine that there are thread sensors on a lot of the spinning and weaving equipment. I will contact them to investigate further.

Regards,

Jason
 
Jason,

Your question about whether there are places to tie intothe machines makes me chuckle. I called Haas once to help me find some I/O points to permantly wire in my robot to a model B I/O card. I really only needed a chuck clamp/unclamp command, cycle start, and cycle complete - to be used as a robot call. The rest of the prox switches etc, I'd wire into the I/O unit directly, since the only time I really cared whether the door was open was when the robot arm had to go through it (or not). Surely I figured they ultimately end up at an I/O point on a main board somewhere, but where was the question.

After hemming and hawing and putting me on hold for 30 minutes the tech came back and said there is a foot peddle that opens and closes the chuck... duh, I asked. Any idea where it ties into? Nope. Next was the cycle start... 30 minutes later he asks if there is a green button on the contorl panel... arrggggh.

Now I knew why our local manufacturer had the robot just push the buttons with a rubber tipped poker on the end of the effector.

I've yet to find a I/O point for remote I/O on any of our Brother or Juki computerized tackers. Realistically though you simply need to paralell into the cycle start button to start the cycle, and perhaps the presser foot or clamps. You can steal the signal from the cycle complete beeper if you have one for the robot call, use the robot timer, or add a microswitch that is triggered by something like the presser foot auto lift at the end of the cycle, etc.

If you find out how to get around the bobbin case issues on a lockstitch machine, let me know. But then, you'll make billions off it.

Could you use an Industrial Serger to sew the seams? That would eliminate the need for a bobbin thread, but they are a lot trickier to keep threads from breaking in my experience.
 
Stuart,

The Tackers that I have do have rotary actuators tied into them so I belive that you can tie other things into it. These are used for a U shape piece that rotates so that you can sew a complete box using the two stage clamping. These guys here look like they have done it:

http://www.tennsew.com/MilitaryAutomatics/MilitaryAutomaticVideos/RunTopVelcroStk.wmv

some processes that I woudl like to do are with an overlocked edge, others might able to be acomplished with a chain stitch.
 
I have automated several sewing machines for production sewing -

1.) We used standard 3/4HP Baldor Induction motors for the sewing machine operating in closed loop vector mode and a small prox switch on the machine (upper part of reciprocating shaft that holds the needle) for homing / stopping with the needle in the up position.

THREAD DRIFT ALERT

MG, can you elaborate on how to do that? I want to be able to stop a vertical slotter in the up position electrically. One slotter is a dedicated machine with a 3 phase motor, the other is a 120V Bridgeport E head. Neither has any mechanical way to stop the slide in position. This is part of a larger project. The E head may be mounted on a VMC, the slotter may be automated with a PLC or with servos.
 
Have you looked at using an embroidery machine? I used to work for Melco and their XT could run at 1500 spm for short stitches in any direction. Sew field is just over 400mm x 400mm.

For quite a bit more money I believe you could get them to add a laser cutter for non-PVC products.

Good luck with being able to control the motion of the piece under the needle. It's VERY difficult to control the tension and feeding to get it to work right. The Melco people spent years on the programming to get it reliable with different threads and materials. It even depends on the humidity and temperature. Design of the some of the specialist pieces like the bobbin case had to be done in cooperation with special suppliers. Thread tension sensing is no fun either.

Chris P
 
It looks like it is just a bunch of spike stabbed into the rug, the rug is sliding on the smooth plate.
When the spikes retract, the slide on the side has grabbed it, to slide it off the table.
That's how it looks to me anyway.
 
Thanks Chris for the recommendation, but this will be an edging process just like in the video. It will be one type of thread and the machine I built for this process has no way of changing the typical sewing machine changes. Once I have the setup down, it should be completely repeatable.

on the video, I think you are right about the spikes. I want to hold in the same way, but I need to use a vacuum for this. The material is extremely decorative and a bunch of holes in it would be seriously bad.

What about some form of electromagnet clamping? Soemthing that can turn on and off with a plate below. Would this be possible?

Thanks

Jason
 
Are you planning on having your material sewn with magnetic thread so an electromagnet can pick it up. Now that would be unique...

With enough vacuum lift points and a supply large enough to handle the leakage vacuum would work.
 








 
Back
Top