Position repeatability?

[Black Forest]

What would be the repeatability on a good quality CNC plasma table? And how do you all measure for it? How accurate are the top of the line CNC plasma tables?

For instance if cutting multiple parts of the same dimensions how close do you all get to being equal?

[Black Forest]

For those of you that nest parts how close are they to being the same size?

Industrial tables using air plasma and not HD are they more accurate? Say a Hypertherm table for instance.

[Gamelord]

If you have a good quality CNC, every piece should be exactly like the first one or the last one. If you have irregularities between two pieces then you either have something loose on your table that is creating slippage or you have a very low quality CNC that can't hold tolerances....which would be worthless to own and probably not very likely.

As for nesting, it depends on what plasma you are using and what thickness of material. The kerf on my Custmaster 101 is appx. 5/32" when cutting 3/4" or thicker at 80amps, 3/32" when cutting 3/8" material at 60amps and about 1/16" when cutting 14ga at 40amps. Normally I leave about 1/8" between cuts if I have to squeeze things together tightly. If not then I usually leave 3/8" between cuts just because the excess gap is less costly than ruining a much larger piece of metal.

Hope this helps.

[Black Forest]

Thank you Gamelord. When you cut several or many of the same part what would your parts be relative to each other? I know you wrote exactly the same but some peoples exactly the same are plus or minus a few thousandths of an inch or mm. Just sayin! I ask because I am building my table at the moment and want it as accurate as possible so I will use ballscrews and rotate the nuts instead of the screw itself. I will use 32mm diameter screws with servos driving the nuts.

[Gamelord]

If your table is exact, your parts will be exact. The only time I will see any measureable differencer in my parts is when the consumables are warn. Measrueable difference would be less than a couple thousandths. If you need more accuracy than that, you should probably go laser or waterjet.

[whiskeymike]

Hard to say, I haven't measured a high number of pieces, but I would guess mine are within about .030-.050 of an inch different. I have a stack of 50 pieces of 16ga stacked up right now and you can feel the ridges, but some of that is light dross and will be cleaned up. You'll find that the smaller cuts like circular holes are more off than others. That's while you'll find suggestions for drilling or punching those out. HD plasma is definitely better. Water Jet and laser are even better. Search for Jim colt threads referencing HD plasma and you should find the accuracy differences.

What accuracy do you find acceptable.

[Black Forest]

What I am trying to find out is what is possible regarding accuracy with air plasma. Is it right to assume that the thicker materials would be harder to get parts exactly the same? Do industrial tables use rack and pinion or ballscrews? I would think the top of the line tables would use ballscrews but I am not sure. Until now I have all my parts cut with laser or waterjet. I am just looking to find out what I could expect for accuracy if I don't skimp on the mechanicals of the table.

[jimcolt]

Industrial cnc plasma cutting tables that are designed for modern high definition plasma cutting (tighter tolerances than air plasma) use helical rack and pinion designs with zero backlash planetary gear reduction as well as inertial matched brushless servo drive systems. Repeatability and accuracy of the machines is incredibly tight......with some manufacturers claiming repeatability to .0005 (1/2 thousandths). Expect to pay over $100k for one of these machines....which are similar in design to the cnc machines used for water jet and laser cutting.

Then...you need to add the accuracy of the cutting process (water jet, laser, plasma (air, oxygen or High Definition) or oxyfuel).....each of which will vary depending of power level, material type and material thickness. You cannot just talk about one number for cut part accuracy for all power levels and thicknesses.

When discussing entry level and light industrial cnc cutting machines.....expect different sets of accuracy numbers.....especially over time as the machines are of lighter construction to keep costs within the buyers budget!


Jim Colt Hypertherm

[Black Forest]

I looked today at helical ground rack and pinion. They are quite proud of the stuff! Do the high end tables use rack and pinion because of the environment or is there something else better about them than ballscrews?

[Brand X]

I would guess that air plasma process causes more error then a decent mid grade table brings into the mix. Same goes with the operator. Careful setup of parameters can make a huge difference. I think a incorrectly setup or poor THC causes more error then most people realise.

I can run a scribe on mine, and see just how good the machine is in general. Kind of impressive really,even with the stylus/etc.. having some slop in it. Gives lowly Joe blow in his shop, a pretty good tool to work with.

[B Holmes]

For what it is worth, I measured some 14 ga. parts today and found a max of .006" variation. Most were within .003".

[Black Forest]

Thank you for the information. I will write what I end up doing.

[msimpson99]

I cheap hobby CNC made from particle board and aluminum V rails, if built correctly, can achieve some remarkable accuracy and repeatably.

This photo was take over 6 years ago. I used a small Vbit connected to a Festool Router mounted on the the above machine.

IMG_0166.JPG

The machine was using 5Start ACME screws and Derlin nuts. My current main CNC machine (KRMx02) uses R&P, and while not the helical kind, has negligible backlash. I thinks its on the order of .0007 last time I tuned the machine. One of these machines (A very small one) was converted to Plasma recently.

_MG_3860.jpg

While I can cut some remarkable things using Air plasma, it can in no way, shape, or form, come close to the accuracy of the machine.

That said, If you start out with machine that is fairly accurate, it will make it easier to dial things in and at the very least, get consistent cuts with good consumables.

[jimcolt]

Ballscrews are expensive. Ballscrews need to get very large in diameter when the length increases in order to combat the twisting effect of torque. As the diameter and length increases....now you must increase the size of the drive motor and inertia matching. There are some cnc plasma machines that use ballscrews....often just on the gantry axis as the length is relatively short (look at Whitney Punch Plasmas as an example).

Rack and pinion drive mechanics are certainly the most common setup...easy to design and reasonably priced.

Jim Colt

I looked today at helical ground rack and pinion. They are quite proud of the stuff! Do the high end tables use rack and pinion because of the environment or is there something else better about them than ballscrews?

[Black Forest]

Jim, do you know of any industrial tables that use ballscrews but drive the nut and not the screw? For sure on a very long table rack and pinion would be the only practical solution. My table will only have a cutting area of 5 x 10 feet so a ballscrew with 32 millmeter diameter is not prohibitively expensive.
The design is such that the linear rails and ballscrews will be protected and where the gantry frame enters the enclosure with the ballscrews and rails there will be either two opposing brush type strips or the rubber sealing strips as used on glass digital linear scales.

[BTA Plasma]

No there are none currently but there will be one shortly. Rotating the nut requires a very nice set of angular thrust bearings and some preload. It is a totally different type of frame build. It takes a very precise process to obtain a machine that will do that and stay within tolerance for decades.

[beefy]

I cheap hobby CNC made from particle board and aluminum V rails, if built correctly, can achieve some remarkable accuracy and repeatably.

My X rails are nothing but 1" angle iron mounted on top of box section tube, the corner of the angle iron pointing upwards. I made my own V-rollers in the lathe and they run on top of the angle iron.

I used fishing line stretched very tight to get the angle iron rail very straight (plus the help of a strong magnifying glass). I've certainly got no problems with accuracy or straightness or smoothness, although I do periodically clean and wax the rails to stop rust and keep them smooth.

Hiwin recirculating bearings are great but at the end of the day they are only as accurate as the surface they are mounted on, and how accurate the X and Y framework is constructed, i.e. are they perpendicular to each other, etc. I think that sometimes gets overlooked and many focus on whoopee doo precision rails.

Once you have a well set up table you're at the mercy of the "floppy" air plasma flame. I think that's the biggest factor for repeatability. I'd heard a few plasma guys say they get all their plasma parameters tuned in then come back on a different day and the cut is not quite the same and they have to retune things.

Some time ago I looked at the thermal coefficient of expansion of steel vs aluminium. From a repeatability (as opposed to accuracy) I prefer steel because it's length changes a lot less from the worst balls shrivelling winter weather to the hottest of the summer weather.

Keith.