[fredd3039]

Hello,
I need some advice on how to build a linear motion platform with single axis cnc control. I am not sure if I even need cnc control for this. I need a 6.5' linear guide rail that can support roughly 75lbs. The purpose of this linear guide is to move a 36" 75lb object on its surface back and forth along the axis. No other axis are needed. I already have guide rails and blocks, plus ball screw and plum connector. I have a: 2X Linear Rail HGR20-200-2000mm 4X Blocks Ballscrew RM1605 BF12/BK12 CNC Set, I purchased off of ebay for another project and it is what I am going to use for this project. I guess my question is what is the best type of motor to use and how to control it. It only needs to move on on one axis back and forth. Any ideas?

[spumco]

Need more info.

Other dimensions/characteristics of object (is it fragile, is it a dumb block of metal, is it jiggly)
Desired maximum speed of move
Ballpark acceleration ( needs to move 6.5 feet in a minute, 10 seconds, 1 second, etc)
Accuracy/resolution requirements
Is this point to point, or need to position somewhere between the ends
How long does this have to last?
Is it in a harsh environment?
How do you want to control it? (PC control or stand-alone thing with no computer attached)

[fredd3039]

Need more info.

Other dimensions/characteristics of object (is it fragile, is it a dumb block of metal, is it jiggly)
Desired maximum speed of move
Ballpark acceleration ( needs to move 6.5 feet in a minute, 10 seconds, 1 second, etc)
Accuracy/resolution requirements
Is this point to point, or need to position somewhere between the ends
How long does this have to last?
Is it in a harsh environment?
How do you want to control it? (PC control or stand-alone thing with no computer attached)

Basically there is a 36" magnetic chuck that weighs 75 Lbs that needs to move from X,0 position to X,36" then back to X,0.
The chock is steel and very sturdy. The chuck will have a 30" long flat piece of steel mounted on its surface. The steel is to be surfaced and finished by a modified belt grinder with a rubber wheel. The wheel will only be removing between .0001 and .001 per pass depending on finish requirements. The accuracy is extremely important from the starting position as this is where the depth adjustments are made and there will be a a radius from the wheel that must be maintained. As far as the end of the metal, the wheel just rides off of the metal at the end of the cycle then returns back to zero. Because of the flexibility of the rubber wheel, even at the max durometer, the wheel may have to make several passes over the surface to obtain the proper finish. This is why returning to the same zero is so important. We must be able to preserve the wheels radius through the various grinding and polishing phases. It is not in a harsh environment, the acceleration is not something I am sure of. When hand cranking a similar setup we can go as fast as we can manually turn the crank wheel so who knows. As far as controlling the machine. I am open to any suggestions.

[spumco]

My uncalibrated brain equates "fast as we can manually turn the crank wheel" equals about 150-300IPM. If you can guess at the time it takes when cranking over the 30 inches let me (and anyone else lurking) know and we can ballpark the feed rate.

So this level of positioning accuracy rules out pneumatics and hydraulics. It can be done with those, but the control system is fussy. First pass: ball-screw, belt, rack & pinion, or linear motor.

I'm thinking you want a stiff drive. Not so much for cutting forces but for repeatability and settling time. So long belts are out unless you get a really stiff & wide one.

Rack & pinion is still in the running if you have a linear encoder on the axis and a low-backlash rack and pinion & gear/belt reduction. Belt drive would need the same things (linear encoder)

A ballscrew that long should be OK, but it'll need to be pretty big (diameter) to deal with the unsupported length. And those are expensive if they are high-accuracy. If you go cheap, then you still need a linear encoder to deal with screw pitch variations (and thermal growth)

Linear motor would be perfect, but terribly expensive if you don't have the time/ability to hunt ebay for surplus.

First spitball: belt drive. Belt would be smooth and quiet, low maintenance, and cheap compared to a high-quality ballscrew. Alignment isn't nearly as fussy as with a ballscrew that long.

Use as wide a belt as you can fit, like 32mm or 50mm. Use a GT2-8mm or AT10mm belt with 20T pulleys (these meet the minimum diameter for both belts). The table drive and idler pulleys are mounted to the frame and the drive end is driven with a servo and reduction gear/belt. Drive & idler pulley shafts are supported at both ends on bearings (no over-hung shafts to flex). A second idler in the middle of the long unsupported belt span would help with belt vibration

You would probably want a 750w-1.5kw servo to drive it, and maybe a 3:1 or 5:1 reduction at the primary. You don't need a servo motor (a NEMA34 stepper would be fine), but for reasons below a servo would be preferred.

Linear encoder mounted along axis. A HiWin or other 1-micron encoder should work... maybe one of the newer field-installable encoder tape things could be stuck on the frame or underside of the table.

Control could be as easy as a 'smart' servo drive. Most of the brand-name(ish) drives have internal memory for indexing functions - you could build a control panel with a few buttons on it that are attached to digital inputs on the drive. Button 1 - home axis. Button 2 - start cycle. Button 3 - jog X-. Button 4 - jog X+. Your entire control system would be:

Disconnect switch
240vac EMI filter
240vac Line reactor
Couple of breakers
24VDC drive logic PS (depends on drive)
Contactor for drive main power
24vdc E-stop relay (to drop drive main power contactor and power any buttons/lights/gizmos you want on panel)
Drive
Enclosure & fan
Control panel. As simple or complicated as you want.

Allen-Bradley, Delta, Copley, and others all make servo drives which can accept aux encoder signals for positional accuracy. If you want all new parts with a warranty, Delta is the cheapest that has aux encoder inputs (that I know about).

If you're bodging something up on a tight budget, an Ultra3000 2098-DSD-020X would drive any servo you want and has aux encoder input plus plenty of internal indexing program slots. I think with simple digital inputs you can choose up to 32 indexing functions.

If you pick a drive that doesn't have internal indexing or programming, a cheap Automation Direct PLC would be the ticket to do the indexing/motion commands to the drive.

Have fun researching all the above - it's all on the internet, BTW.

-Ralph

[datac]

>>>>> I am not sure if I even need cnc control for this.

A really, really cheap way I have been replacing "hand cranking" with a semi Automatic non-CNC solution is via these:

https://www.amazon.com/gp/product/B0...?ie=UTF8&psc=1

I have used these on a mill and a lathe so far and frankly, love them. They have a few different "run modes", but very easy to control feed rates and easy to locate limit switches to force it to automatically reverse at each end of travel.

At minimum, it can be an easy to wire up "Test Controller" used to prove out your mechanical system.

[spumco]

Datac - that looks like the best thing for a test bench I've seen in a while. Thanks for the link.