[Hi-Vac]

Hello everyone,
This is my first post to this forum. I have been reading for awhile as an anynomous user. I am very impressed with the knowledge out here. Anyway, back to my question.

I am in the process of building a 4'6" X 10' Robotic plasma table. See attached photo's. I have built the table, and we are in the process of specifying the software / hardware to use to control it. I am leaning towards Camsoft with an Ethernet controlller card. (Camsoft Professional Software - 6 Axis DSP Ethernet Card).

But, I am still having trouble deciding whether to contunue the robotic process or scrap the robot idea all together and build a one sided gantry system to slide along the back instead. My concern is the accuracy of the slider.

Also, I am assuming that auto height control will be implemented with the robot.

Has anyone ever attempted something like this?

Does anyone have any concerns suggestions on camsoft or a replacement?

I can provide any additional data if needed (motor sizing, robot type, etc...)

[Al_The_Man]

I see you already have the robot, ASEA by the look of it. My spin on it would be that a Robots has its place, especially if you want to cut parts out of a preassembled shape for example and do all kind of vertical & horizontal cuts etc, but if you are doing all the cutting on a horizontal surface out of plain sheet, a robot is a bit of overkill, You will probabally need AHC whether a robot or gantry is used, were you thinking of designing that also? Why does it have to be a one sided gantry? Also would'nt 4ft6"x10" be stretching the work envelope of the robot? ( although I took a second look and it appears it also runs on rails!). Were you looking at implementing a teach function on the robot? I would say Camsoft professional should handle either robot or gantry, although with the rail function I can see the robot being pretty program design intensive . Do you have the amplifiers for the motors on the robot?
Al

[Hi-Vac]

Thanks for the reponse. I actually have two of the Asea L6/2 robots. One is on a 5' slider. This allows us to cut approx 42" out of an 48" X 8' sheet of metal. Hindsight, I would like a slightly longer reach on the robot, but it will handle all we need at this time.

I agree the robot is more than we need and I know this should not be a design factor..... but they look really cool running!

Yes I have the control cabinets as well, but I was not planning to use the existing amplifiers. I may change my mind after I get the costs for new ones.

As far as the one sided gantry (if we decide this would be better) goes, I would like to keep the front of the table open for loading / unloading parts.

[Al_The_Man]

As far as the one sided gantry (if we decide this would be better) goes, I would like to keep the front of the table open for loading / unloading parts.

Most plasma tables have the X axis travel mechanism below the top of the table.
With the gantry at one end this allows clear edge loading of material etc.
Also It looks like you have some kind of tilt mechanism on the table? Also with a plasma table in a production environment you need a practicle way to remove the scrap that falls to the bottom of the table, unless you have already thought of that.
Al

[balsaman]

Are you sure you can control the robot via regular gcode? Most use propriatary software and need to be taught. Those robots are ABB no? Still parts available for them? If so they will be expensive.

An interesting project.

Eric

[balsaman]

BTW, robots need very expensive cages and safety circuits around them to be legally run.

Eric

[Hi-Vac]

Thanks,

Actually, the table doesn't tilt, the table has ball transfers that will pop up to facilitate material handling (See photo). The angle of the table is configured to allow slag to "slide" to the front for easy cleaning.

Yes they are ABB Robots. And yes, I am sure the repair costs are high. I have found several vendors that sell replacement (or rebuilt) servo motors for this exact model.

I am told from Gary @ Camsoft, the programming is embedded in his software. additionally, Gary expressed that the professional version will allow importing of .DXF files (from autocad).

Obviously, I am new to CNC, but I enjoy unusual projects. I own a small business that manufacture's environmental equipment, and need to be able to cut parts projects. I could have just bought a CNC cutting table, but I wanted the challange. I am at a stage in my life where this is more fun.

BTW: The cost is not a real issue.

As for the safety issues, I am aware of the need for a laser curtain around the working area of the robot, but I was not planning to inclose it in a fence, maybe some of these issues will push me to a gantry instead.

[unterhaus]

EMC (www.linuxcnc.org) has the kinematics for a robot. I am sure there would be some work to be done to get it working again with the most recent revision, but I think the guys at NIST would probably help. The advantage would be that you could use G-code.

Someone recently was going to use Mach 2 on a cartesian robot by inverting the kinematics.

I have a GE robot which I took apart for parts. The amplifiers for all 5 motors were on one board. If your robot has discrete amps, I would try to use them. They would probably be analog command input which would mean either the Rutex step/dir to analog converter, or one of the many analog output options that works with EMC.

[greggv]

I would definately look at a Delta Tau controller before Camsoft. A lot more configurable, and a lot more features, and, I think, a better NC innerface.

[Hi-Vac]

To everyone, thanks for all the input. A question rwas brought up ecently in our office was the feed rate (ipm) of the robot movement ( I know... This should have been considered before now!),

We are looking for 300 ipm range.

I am thinking again about a cantilevered gantry (one sided), and based on this option, I am considering Rack & Pinion movement ( for the x axis ). What would be a good choice for the slide (example 3 or 4 thompson rods, roller bearings, etc) The x axis will need to traverse 10'

[Al_The_Man]

I would look at the plain linear bearing & track by Pacific Bearing for the outer rod and maybe an inner track with a cam follower style bearing that would have a front & rear sacraficial scraper for the track.
Plasma's put out a phenomenal amount of metal dust into the atmosphere, which brings up another subject of a dust extractor system of some kind.
Most of the commercial systems extract from under the work piece which means the table requires temporary covering around the sheet being cut.
300 ipm is usually the top end of the cutting speed (if that) with 1000 ipm for the rapid moves.
Al

[Pythagoras]

It seems to me that the table you bought is asking for a conventional gantry to be put on it. Supported and driven from each side. Either by dual servos or a drive shaft.
Why reinvent the wheel and make it more complex than it has to be by introducing a robot into the equasion. The question you need to ask is, do you want a machine that works well, or an industrial peice of artwork? To go down the conventioanl path would be comparitivly cheap and easy by comparisom to using the robot. The thing is I know you can achive success using the conventional method, where Using the robot could cause all sorts of head aches.
Because I have a similar machine to what you could have, I can tell you that it shouldnt cost more that a few thousend bucks to get set up. Keep this in mind when comparing the cost of getting your robot setup up to scratch.
Keep in mind also that you could use cheaper off the shelf software with a conventional setup also.

[Hi-Vac]

Thanks everyone for the help in making this decision easier. I understand the complexity of the robot, not to mention the added issue of the gantry, the limited reach (cutting path), but the deal breaker was the movement speed. The Asea robot moves @ 1.3m/s Which (I Think I am right?, see below) is about 30 in/min. without varying the amperage of the plasma unit (which I haven't purchased yet), this would prevent thinner materials to cut cleanly.


Here is the specification on the robot:
( I don't really understand the second or third reading... is it millimeters, meters, and where is this measurement taken? I have no problem with the degrees per second.)

ASEA IRB6/2 Industrial Robot
1. Base rotation has a working range of ± 180° and a max axis speed of 114° /s
2. Radial arm motion has a working range of 80° and a max axis speed of 1.3 m/s
3. Vertical arm motion has a working range of 65° and a max axis speed of 1.3 m/s
4. Wrist rotation has a working range of ± 90° and a max axis speed of 138° /s
5. Wrist rotation has a working range of ± 180° and a max axis speed of 234° /s

So ..... Gantry it is then. Now, I have to design this, figure out what speed, ratio's, motor size, etc.... oh joy, at least I have a "great" resource in CNCzone.com!!!!


Thanks again everyone!


John....

[unterhaus]

the deal breaker was the movement speed. The Asea robot moves @ 1.3m/s Which (I Think I am right?, see below) is about 30 in/min. .

1.3m/s is more than 3000 in/min

The reason that these robots are so dangerous is that they can put a hole in a person because they are moving so fast. Speed would not be your problem, although you have probably made the right decision anyway.

[Hi-Vac]

Hello everyone.

Last week when we last spoke………ok, Ill admit it, it been a little longer than a couple of weeks! More like 4 years! (You know the drill, new marriage, new daughter, business is “off the charts” busy, etc….well at least the first two are correct )

Anyways, after a several year hiatus, I have finally revisited our plasma table design. As you may (or may not!) remember, we originally built this table to encompass a gantry mounted robotic arm, witch we have since abandoned, and now we have this awesome table with no gantry! Our decision to create a one sided gantry arm has met a little resistance in the fact that the x axis moving weight would be close to 500 lbs in the current configuration as depicted in the sketch attached. Apparently the “inertia” or the resistance to accelerate the 500 lbs seems add some additional problems in order to create and provide clean cuts. Here is some “ideas” we were tossing around


X Axis:
• 4 - 1.50 Thomson Twin Pillow Block Bearings mounted 24” apart in both directions.
• Approx 106” of X-Axis travel (122” overall) driven by a 1.5” x .500 lead anti backlash ball screw (Direct Drive).
• The gantry would be constructed using two ¼” plates of steel laser cut and welded together with 12” separation plates, thus creating a boxed design. (for the records, all of the pieces would be notched and the corresponding plates would have matching slots, thus making assembly and alignment a snap) The main side plates would be cut from one plate of steel.
• Drive system would be either a DC or possibly AC servo motor with encoder.
• The entire linear motion, including the ball screw and rails would be covered with one rectangular bellow.

Y Axis:
• Two 1.5” Thomson Twin Pillow Block Bearings mounted one on top of the other with 9” spacing, riding on two supported shafts mounted on the sidewall of one side of the main gantry plates.
• Approx 60” of Y-Axis travel driven by a .7500 x .500 lead anti backlash ball screw (Direct Drive)
• Drive system would also be servo with encoder
• The entire linear motion, including the ball screw, would be cased inside the boxed gantry.

Z Axis:
• Two THK 12 series square slide bearings 350mm long, giving us approx 4.5” of z-Axis travel using a .500 ball screw with a non preloaded nut. This would be timing belt driven.
• Hopefully, the software we choose allows for THC.
• The entire linear motion, including the ball screw, would be cased inside the boxed gantry.


Some of my questions are as follows:
1. Should we be concerned with the weight of the x axis? I was told the limit for the gantry should be under 150lbs in order to get the acceleration needed to make clean and accurate cuts using DC servo’s.
2. Should we revisit the gantry design and possibly make it lighter using aluminum plate instead of steel (saving almost 200lbs) or should the answer be larger AC servo motors.
3. Should we ditch the ball screw and go with rack and pinion? (definitely, not my favorite) on the x axis?
4. I like the 3 axis package designs available from some of the software suppliers that provide you all of the necessary equipment including the servos and encoders, but is this really the right way to go? I would like this system to only be a controller at the table, and all of the CAD / Cam work would happen “upstairs”.
5. It has been suggested that we ditch the one sided gantry and “add” a rail to the front side of the table, Although I an some way agree this would be a good solution, I would rather exhaust “ALL” options for one sided before we “concede defeat” to a two sided gantry design. It mostly has to do with the way we will use this system.
6. What about this “touch and go” regarding the piercing of each cut. I was told in addition to THC, we would want the z axis to come down and touch off the metal, re-establishing the z-axis zero plane, each and every time we start or pierce a cut. I was under the assumption, that initially, we would “program into the software” the thickness of the steel sitting on the table, and this would provide the system with a new “z axis zero plane”. And the THC would take care of the rest, but now I am told we should set up our z axis with the torch on a spring or shock mounted slider (additional to the original z-axis slider) that has a limit switch to detect the plate during this “touch and go”. What are your thoughts?

Obviously, the table is complete, and the bearing and supported shafts are selected, but the rest is “still” in cad. So any and all ideas / changes would be the easiest to implement now, so I would love to hear your thoughts!


Thanks,
John

[WSS]

Hi-Vac,
The thomson ssuo-24 series will work with the weight you are speaking about. I have a set of 1.5" open bearings with over 8 million travel inches in harsh environs. The load on them is 400lbs hinged with a 12" gap. I had to float the inverted side to get in line with the tight tolerance. I have had one bearing fail due to some carbide that went through the circulating channel. It was instant and easy to find when it happened. I can export that section of the drawing into a jpg if you think it is of interest. One thing you have to take into consideration is the addition of wiper seals, they add more friction than I imagined. You would likely have to factor that into your motor torque needs. No where do they list the specs with the wipers (at least 10 years ago). Without the wipers it is like stepping on a skate board. 500lbs you could move with your pinky, literally. The machine We have is low speed compared to CNC plasma, We top out at 30ipm, we are not accelerating near what an efficient plasma needs. So the starting torque may be your limitation ($$$).

WSS

[Hi-Vac]

WSS,
Thanks for the reply. Yes, I would be interested in your "inverted design" if it's not too much trouble. When you say limitation ($$$), due you think the acceleration could be achieved with more money?

We were tossing around the idea all day of going with a driven nut assembly as offered by THK (see attached PDF). our assumption was that this would give us the abbility to go to a smaller shaft with a shorter lead and still obtain the speed needed (300ipm) for thin sheet plasma cutting. because we were not limited to the maximum shaft RPM's, we could spin the nut fast enough to give us the speed. however, the problem we noticed with the rotating nut offering from THK, is it only comes in a one to one ratio. For example, if you buy a 20mm shaft, it will have a 20mm lead. I think we could "reduce" this to a more managable ratio as simply as the selection of the timing pulleys used, but would it be enough to provide us with the acceleration needed. Also, nowhere does it give us the backlash specs on this assembly. They talk about preload, but I think this is just in the axial bearing assembly, and not in the ball nut itself.

Any thoughts?

[WSS]

HI-Vac,
Sorry it took so long to reply! I had archived the drawing and it was on a CD offsite. It was driving me nuts. Well, I found it. I am not real good on the electronic side , so I really couldn't say if you could get the right motor combo for the app you are needing. You could use the round rails though and build a floor mount gantry to slide your table inside of? Those bearings are a small fortune. I would try to use them if possible.

Good luck, I will keep an eye out on your progress, it looks interesting.
WSS

[Hi-Vac]

WSS,
Thanks for the inverted Thomson bearing drawing. When I pulled it up, it was amazing how similar your drawing looked like mine! I was doing the exact same thing. Inverting the rear (right) bearing as an open bearing of this type's rating was reduced by over 50% if the bearing was pulled up instead of it being pushed onto the rail. I guess this means I am at least 10 years behind you! Thanks for taking the time to show us the drawing.

We have decided to go with a double sided gantry..... I know…. it's the right thing to do.

Interesting thing to mention: We were still having problems overcoming the long axis weight and a fine lead ball screw would accomplish this, but we couldn't spin it fast enough to get the cutting speeds (because of ball screw whip) we needed. If we went with a course lead, we wouldn't get the torque needed. It was a catch 22. So, we flipped the design so the gantry is now straddling the long direction of the table, and the weight is being carried by the two short ball screws (the 4 foot direction) and the long axis (the 8 foot direction) is mounted to the gantry, and is light, so the course lead screw has enough torque and didn’t need to spin faster than the design specs would allow.

After we started looking at the design turned this way, it solved our access to the table problems (which was the same way the one sided gantry helped us), and all of the weight, speed, torque numbers all added up. We decided this design was the way to go for us over the one sided gantry. However, after all that, we then decided to scrap the ball screws, and are going with a helical rack and pinion set-up (supplied by Atlanta Drive Systems – a German Company) with a 39 to one gear box with less than 5 arc minutes of backlash. This gives us (in a perfect world!) a maximum of .0025” of backlash in the drive system. We couldn’t accomplish this with a ball screw without going to some type of pre-loaded nut. The helical rack and pinion achieves this backlash without using the “pressure” set-up. In case anyone is interested, the cost for a 2 meter rack is $262.00 and the pinion is $159.00 (www.atlantadrives.com). They have a USA office in New Jersey.


In case anyone is interested, we scrapped the “spin the nut instead of the lead screw” idea because there was no easy way to “add” a second nut to this configuration for pre-load in the THK’s – BLR line , and if you went with the DIR series, it came with a preloaded nut, but only in a precision ground lead screw ($$$$$$). Without the preloaded nut, the backlash was in the .015” range.

PS:
WSS, We still plan to use the bearings!

[WSS]

Hi-Vac,
That is a neat link you listed. I assume you will use the split helical pinion?

http://www.atlantadrives.com/zerobacklash.htm

Clever design. The site says that it is a spring pack loaded split pinion. I wonder how the rack would wear? That could be integrated into almost any R&P set-up. It may be worth a try on one of the machines we use with the Thomson SSU bearings. I need to replace a pinion gear in the next month or so as it is. I am going to contact them and see what it would cost to integrate on ours.

Thanks for the info!
WSS