Accuracy on the Z axis with CNC
I have two shoptask machines. One with factory CNC setup and one all natural. I am in the process of converting the all natural version to CNC and wondering what type of accuracy I can expect?
My other machine has the ball screw on the Z axis to drive it. With this one I'd just stick with the quill worm drive.
Also, if anyone is interested, I am really modifying my shoptask. I've removed ALL manual controls, motors and extras. Replaced both motors with 2 HP 3 phase VFD big boys. No pulleys anywhere. I'm using gates sprockets with idler/tensioners. All controls are external. In the control box is a computer, transformers, filters, coolant switch, drivers, HVAC relays, VFD, pulse generators, breakout boards, relays etc..... Basically, I'm putting way too much time and money into a small machine... but... I'd rather spend the time / money learning all these on a machine that doesn't cost me two arms and some feet to upgrade to all these goodies. The end goal is to have a completely CNC 3 in 1 machine with encoders, spindle control, coolant, HVAC all controled via controller. Ridgid tapping and threading.....
Any advice you can give me about the Z axis would be appreciated.
Z axis cnc
If your gonna Upgrade the manual machine to CNC (other than the kits from shoptask). then your accuacy may be alot better than the "Nook industry" rolled ball screws. The Nook industries are rated at 0.004/foot.
Having said that, on the Z axis, if BEFORE you mount your ball screws, and AFTER your remove all the mechanical off the Z head, EXCEPT the quill counter balance sping and pinion/rack drive.
Set the spring to counter the "Wieght" of the quill, until when you release it at mid position it will creep back up do to the spring tension. The reason I do this is since the quill is always pulling up against the ball screw this eliminates back lash. In the small travel that the Z will actually travel (that would include your home switch). My realistic Z travel is 3" so in actual practice, I really don't see any inaccuracies in this small of a travel. My Mytoyos calipurs can measure to 0.0001" and when I measure a Z axis hiegth. The average I have seen is about a 0.0000-0.0003 differences. Having said that, I have yet to run into a job that acually NEEDS tolarances lower than 0.001-0.005".
hope this helps,
That really helped me, thanks. So the Nook industries isn't really accurate?
Also.... what if... I let the ball screw come all the way out and the ball bearings fell out.... just what if..... how would I reload the bearings and load the screw. Any web sites that have such instruction?
Do you put any collets/tools in the quill before you test the creap factor? This is with all the controls on and fine feed off right?
Yea, dont do that, you can put the little balls back in but it is a major pain, assuming you find all of them. the Nook's have a little feed tube that you could load them into the nut, and then the tube and then remount.
The Nook industries rolled screws aren't to bad at 0.004"/ft, they are adaquite for the shoptask. Personally under CNC control, I can usually get to 0.001 or a little under on parts depending on its size. I usually work with pieces (like paintball markers, and other parts not much more larger than 10" in overall size, an not higher than 6" max), so things in that envolope can be made very accuratly.
You could goto more expensive, (alot more expensive), ground ball screws and double nuts or preloaded nuts. But the truth is, if your gonna go that far you might as well have your X and Y ways and saddle Commercial ground, buy a proffessioal machine way rebuilder. (Unless you really, really know how to hand scrape and make sure all things are scraped square on the saddle as well.
Originally Posted by borrisl
Or you could invest the money in some decent measuring equipment and determine what the actual error is on your individual ballscrew, then simply adjust your software values for that.
Originally Posted by PoppaBear10
The beauty of rolled threads is that the per length error tends to be very consistent over the full length. If you measure .004 error in one foot, you can pretty much bank on it being .008 over two foot length. Simple to just change the values in your setup files.