Ballscrew error compensation.
- dangercraft
- Offline
- Premium Member
-
Less
More
- Posts: 84
- Thank you received: 0
23 Aug 2012 21:40 - 23 Aug 2012 21:41 #23594
by dangercraft
Ballscrew error compensation. was created by dangercraft
Hello everyone,
I am under the impression that if I use a rolled ballscrew zero backlash double nut preload setup and then use linear scales on the joint that I can setup linuxcnc to correct for any lead errors in the screw? I'm thinking that if this is the case I could use say a Rockford rolled C5 ballscrew and get as much accuracy as the linear scale can provide? Am I correct? This is for my Tree VMC
Frank
I am under the impression that if I use a rolled ballscrew zero backlash double nut preload setup and then use linear scales on the joint that I can setup linuxcnc to correct for any lead errors in the screw? I'm thinking that if this is the case I could use say a Rockford rolled C5 ballscrew and get as much accuracy as the linear scale can provide? Am I correct? This is for my Tree VMC
Frank
Last edit: 23 Aug 2012 21:41 by dangercraft.
Please Log in or Create an account to join the conversation.
23 Aug 2012 22:08 #23597
by BigJohnT
Replied by BigJohnT on topic Re:Ballscrew error compensation.
An encoder on the ball screw will be easier to set up. You can make a compensation file for each axis that contains a triplet for each position as outlined in the
AXIS
section of the INI chapter.
That being said if your linear scale is accurate then the accuracy of the ball screw seems to be a moot point.
What kind of accuracy are you trying to get?
John
That being said if your linear scale is accurate then the accuracy of the ball screw seems to be a moot point.
What kind of accuracy are you trying to get?
John
Please Log in or Create an account to join the conversation.
- dangercraft
- Offline
- Premium Member
-
Less
More
- Posts: 84
- Thank you received: 0
23 Aug 2012 22:30 #23599
by dangercraft
Replied by dangercraft on topic Re:Ballscrew error compensation.
Hey John,
The linear scales have 1 micron resolution and +-4 micron cumulative error. The encoder already on the ball screw is 2500PPR quadrature. I would run both. I am replacing all the bearings as well. The current screw is a ground ball screw but its REALLY worn, though mostly in the center part of the travel. I have only checked the X screw, still have to check the Y and the Z. The screws are all the same 1.5" in diameter.
I am looking at two options, though I am certainly open to suggestions: The first is having the current ballscrew reground, which I am not to keen on doing because of the freight from Panama to the U.S. and back. The other is buying a Rockford rolled ballscrew with a preloaded double bearing setup.
As far as accuracy is concerned, I'd be ecstatic to be able to hold -+.002" tolerances on it, and though I am not holding my breath I think I might be able to do a little better if I take my time and do it right.
Frank
The linear scales have 1 micron resolution and +-4 micron cumulative error. The encoder already on the ball screw is 2500PPR quadrature. I would run both. I am replacing all the bearings as well. The current screw is a ground ball screw but its REALLY worn, though mostly in the center part of the travel. I have only checked the X screw, still have to check the Y and the Z. The screws are all the same 1.5" in diameter.
I am looking at two options, though I am certainly open to suggestions: The first is having the current ballscrew reground, which I am not to keen on doing because of the freight from Panama to the U.S. and back. The other is buying a Rockford rolled ballscrew with a preloaded double bearing setup.
As far as accuracy is concerned, I'd be ecstatic to be able to hold -+.002" tolerances on it, and though I am not holding my breath I think I might be able to do a little better if I take my time and do it right.
Frank
Please Log in or Create an account to join the conversation.
23 Aug 2012 23:07 #23601
by BigJohnT
Replied by BigJohnT on topic Re:Ballscrew error compensation.
Hi Frank,
From all the conversations I've read over the last few years I've come to the conclusion that if your ball screw is worn then trying to add a linear scale to improve accuracy will just lead to a race condition between the drive and the linear scale. With any kind of slop in the ball screw it will be impossible to hold the axis still with cutting forces applied. But then I might be full of crap and not know it... while this is blind sight from quite a distance away if possible I would replace the screw and ball nut with a new one if that is more cost effective than rebuilding your current ball screw. And in my second thought I might just fix what needs fixing and if the X axis is more accurate on the ends I might mount my vise out there for a while if the application allows and focus on getting it up and running... but that is just me.
John
From all the conversations I've read over the last few years I've come to the conclusion that if your ball screw is worn then trying to add a linear scale to improve accuracy will just lead to a race condition between the drive and the linear scale. With any kind of slop in the ball screw it will be impossible to hold the axis still with cutting forces applied. But then I might be full of crap and not know it... while this is blind sight from quite a distance away if possible I would replace the screw and ball nut with a new one if that is more cost effective than rebuilding your current ball screw. And in my second thought I might just fix what needs fixing and if the X axis is more accurate on the ends I might mount my vise out there for a while if the application allows and focus on getting it up and running... but that is just me.
John
Please Log in or Create an account to join the conversation.
- dangercraft
- Offline
- Premium Member
-
Less
More
- Posts: 84
- Thank you received: 0
23 Aug 2012 23:28 - 23 Aug 2012 23:30 #23603
by dangercraft
Replied by dangercraft on topic Re:Ballscrew error compensation.
Hey John,
I'm not planning on using the existing worn ballscrews. I'm debating whether to have the current precision ground ballscrew assembly rebuilt to new condition or to swap it out for a high quality rolled screw like a Rockford unit (though I haven't heard from anyone that has used there products... at least their website makes them look good.) Unfortunately, I suspect that with the shipping, regrinding the ballscrews would cost the same as new rolled screws. Since rolled screws, because of the process, are inherently less accurately formed than ground screws, I think that by adding the linear scales with the new rolled screw, I can correct for any lead errors in the rolled ballscrew. I'm not trying to compensate for wear or backlash, though you're right the lash creates a "floating" condition where the controller would keep oscillating back and forth within the worn area.
Frank
I'm not planning on using the existing worn ballscrews. I'm debating whether to have the current precision ground ballscrew assembly rebuilt to new condition or to swap it out for a high quality rolled screw like a Rockford unit (though I haven't heard from anyone that has used there products... at least their website makes them look good.) Unfortunately, I suspect that with the shipping, regrinding the ballscrews would cost the same as new rolled screws. Since rolled screws, because of the process, are inherently less accurately formed than ground screws, I think that by adding the linear scales with the new rolled screw, I can correct for any lead errors in the rolled ballscrew. I'm not trying to compensate for wear or backlash, though you're right the lash creates a "floating" condition where the controller would keep oscillating back and forth within the worn area.
Frank
Last edit: 23 Aug 2012 23:30 by dangercraft.
Please Log in or Create an account to join the conversation.
24 Aug 2012 11:07 #23613
by BigJohnT
Replied by BigJohnT on topic Re:Ballscrew error compensation.
If rebuilding your ground ball screws back to original spec cost the same as a rolled screw that you would have to add a linear scale to and all the hassle of getting that to work I'm a bit confused but that happens more and more as I get older.
To get from Miami to Panama doesn't seem too expensive unless there are some crazy duties to pay.
www.delvi.com/Home.aspx
John
To get from Miami to Panama doesn't seem too expensive unless there are some crazy duties to pay.
www.delvi.com/Home.aspx
John
Please Log in or Create an account to join the conversation.
25 Aug 2012 01:24 #23646
by andypugh
Replied by andypugh on topic Re:Ballscrew error compensation.
dangercraft wrote:
I think you will probably get that with rolled screws.
I can think of very few situations where the cumulative pitch error of rolled screws really matters. Whilst you might want a bearing housing to be 3" +/-0.001" you very rarely see a part called out at 36" +/-0.001"
And, as JT pointed out, if you do need that, you can correct for the pitch error in software.
For reference:
Is backlash and error compensated leadscrews (i.e., not ballscrews at all). I think it took a day with a laser measurement machine.
As far as accuracy is concerned, I'd be ecstatic to be able to hold -+.002" tolerances on it
I think you will probably get that with rolled screws.
I can think of very few situations where the cumulative pitch error of rolled screws really matters. Whilst you might want a bearing housing to be 3" +/-0.001" you very rarely see a part called out at 36" +/-0.001"
And, as JT pointed out, if you do need that, you can correct for the pitch error in software.
For reference:
Is backlash and error compensated leadscrews (i.e., not ballscrews at all). I think it took a day with a laser measurement machine.
Please Log in or Create an account to join the conversation.
Time to create page: 0.299 seconds