Plasma/laser machine accuracy

[memphisjed]

Boss men are looking at a really nice plasma, and a few knock off brand lasers.
The plasma and lasers both run on semi exposed rack and pinion, servo thru gear head, encoder on servo, not on rack. The plasma, controlled automation - yasakwa servos has positional accuracy spec of .015 - .01 repeatedly. Lasers all unknown controllers ( guessing skinned Mach from screen shots and cabinet builds) running unknown servos state .003 inch positional accuracy, .001 repeatability.
Is this possible with exposed rack? 20 foot bed, so thermal and floor shift would be way more anyway- but even ideal I can not see rack drive giving that accuracy. Am I missing something or is it sales pitch numbers no one will check?

 

[DDoug]

I would be concerned with the whole system making the part, not just one aspect (the rack gear).
Jim Colt at Hypetherm IIRC has written on the subject, and used to post here, but I think he has retired.

 

[memphisjed]

I would be concerned with the whole system making the part, not just one aspect (the rack gear).
Jim Colt at Hypetherm IIRC has written on the subject, and used to post here, but I think he has retired.

I am not so concerned with rack drive as questionable specs. Smoothness/acceleration are much more important, at least in thermal cutting. I have seen belt drive stepper plasmas make decent cuts; linear motor lasers make one toothed beaver chewed cuts. Laser is half thermal and half erosion*, does that change machines priority of movement?
* Zanner’s book on metal coloring explains this the clearest, in school it shown as blue light electron ejection (erosion) generating heat.

Agree, learned a fair bit reading Jim’s posts and videos scattered about the internet.

 

[Freedommachine]

Boss men are looking at a really nice plasma, and a few knock off brand lasers.
The plasma and lasers both run on semi exposed rack and pinion, servo thru gear head, encoder on servo, not on rack. The plasma, controlled automation - yasakwa servos has positional accuracy spec of .015 - .01 repeatedly. Lasers all unknown controllers ( guessing skinned Mach from screen shots and cabinet builds) running unknown servos state .003 inch positional accuracy, .001 repeatability.
Is this possible with exposed rack? 20 foot bed, so thermal and floor shift would be way more anyway- but even ideal I can not see rack drive giving that accuracy. Am I missing something or is it sales pitch numbers no one will check?

Shop I work for has a twin spindle gantry mill made by quickmill. It has 100" of Y axis travel across the gantry where the spindles are driven by ballscrew. The gantry itself has 180" of travel and is driven by gear rack. It uses 4 red top fanuc servo motors to move the gantry and take up backlash. Positional accuracy is .002" and repeatability is 0.0005" (checked with ball bar and laser every year). However, the gantry does have 180" scales mounted on either side to track position as well as servo mounted encoders.

Most plasma systems are set up using the system you described. Open rack, closed loop servos with motor mounted encoders. They can be surprisingly accurate when the system is designed properly.

 

[dana gear]

We still use two Air Plasma CNC machines that utilize open gear racks and Anemics start servos with internal encoders, very accurate. These tables have thousands of arc on time with no real problems to report at all. Even with thin material and high travel speeds. There's a lot of second level CNC tables out the with just about every problem imaginable, that said the difference between the good and the bad is day and night. Ball screws around plasma dust is the kiss of death, belt driven has their problems as well, maybe not so much at first but overtime it will show up in most cases. Protected gear racks, proper gear pitch/width, good quality direct servo drives seem to work the best.

 

[Strostkovy]

We've run all sorts of systems ranging from exposed straight cut racks to enclosed helical racks to ball screws. The cutting process was always the limit. The cheaper rails and rack were better than the plasma cutter, and the nicer rack and rails were better than the laser kerf consistency (CO2).
I think any well built machine will work fine as long as the construction appears to be well though out.

 

[GiroDyno]

I've used laser trackers to do volumetric compensation on several 5 axis mills with travels ranging 80 to 100 feet (not inches!).
Some ran on linear motors with scales and some on exposed rack and pinion with encoders on the motors. The rack and pinion with motor encoder machines were capable of hitting +/-0.020" over the entire envelope in any single axis without any compensation beside what was in the Fanuc controller. We could cut that in half running all 5 axis after a comp.

I'd imagine (at 20') if the machine is built with reasonably spec'd components and a sturdy frame on a good foundation it should easily be able to cut +/-0.010" or better.

Probably a safe bet that the 0.001"<0.003" spec on the laser is just a number ripped straight off the label of the no-name motor, and does not indicate the spec of the complete machine... Or its just completely made up!

 

[BT Fabrication]

a new machine should be positional accuracy within 5-10 thou. worn can be all over the place.
laser kerf can be 0.008" wide or plasma can be 0.030" wide then depends on the parts and if they move while being cut.