CNC improvements


Page 1 of 3 123 LastLast
Results 1 to 20 of 48

Thread: CNC improvements

  1. #1
    Member
    Join Date
    Nov 2014
    Posts
    75
    Downloads
    0
    Uploads
    0

    Default CNC improvements

    Hello friends!I am in need of your help if this is possible.I have built a momus cnc that want to use it for luthiery projects and general woodworking. I have some accuracy problems.I am trying to cut a 50x50 mm square and it cuts 49.4 x 49.7,The steps per unit if changes will cut it correct but then on a bigger item they will cut it oversize....What can i do to improve the accuracy(the repeatability is good). the pulleys i have are offcentered is this causing the problem?I cant find pulleys with 8 mm bore xl and also 10 teeth ...the ones i used where opened in a lathe from a 6 mm bore to 8 mm but are 0.2-0.4 offcentered.....Also the cables tha i use for the motors ares simple cables not ground ones....any thoughts?Maybe if i use another type of belts and pulleys?

    Similar Threads:


  2. #2
    Registered vtxstar's Avatar
    Join Date
    Aug 2006
    Location
    Anderson, SC USA
    Posts
    132
    Downloads
    0
    Uploads
    0

    Default Re: CNC improvements

    There are several variables to your questions.

    What Software are you using?
    What motor drive are you using?
    What is the runout of your spindle/router?
    Are the parts uniform? i.e. is the 49.4 x 49.7 square and uniform, or does it vary and the part ends up trapezoidal shaped?
    Is the error repeatable or does error vary?

    True, it would seem that a non-concentric pulley would seem to introduce a variance in position, but the variance should be repeatable over the range of movement.
    Many possibilities exist.

    -=Doug

    "IT ≠ IQ " Starwalt 1999


  3. #3
    Member
    Join Date
    Nov 2014
    Posts
    75
    Downloads
    0
    Uploads
    0

    Default

    Quote Originally Posted by vtxstar View Post
    There are several variables to your questions.

    What Software are you using?
    What motor drive are you using?
    What is the runout of your spindle/router?
    Are the parts uniform? i.e. is the 49.4 x 49.7 square and uniform, or does it vary and the part ends up trapezoidal shaped?
    Is the error repeatable or does error vary?

    True, it would seem that a non-concentric pulley would seem to introduce a variance in position, but the variance should be repeatable over the range of movement.
    Many possibilities exist.

    -=Doug

    Hello i would like to asj somethibg else...the first problen was solved it was a faulty stretched belt.But i stikl need more accuracy tha 0.2 mm... i am thinking of changing the belts with this kind of lead screw http://m.ebay.co.uk/itm/3D-Printer-8mm-Lead-Screw-Rod-300mm-4-Start-Z-Axis-30cm-Linear-Rail-Bar-Shaft-/191382736803?nav=SEARCH&varId=490496660274 and an antibackshlash nut. Whats your opinion on this?thanks!if others can help please do!



  4. #4
    Member awerby's Avatar
    Join Date
    Apr 2004
    Posts
    5728
    Downloads
    0
    Uploads
    0

    Default Re: CNC improvements

    Looking at the design of that machine: momus CNC | benchtop DIY router plans | home page it's not surprising that it was not very accurate. The idea of using a long timing belt instead of a leadscrew isn't that great for CNC cutting machines, although it seems to be a popular way of driving 3D printers. But printers don't have to contend with cutting forces. Replacing that system with leadscrews should help, although it's hard to say how much, since no specifications about the accuracy of those screws were included in that ad.

    [FONT=Verdana]Andrew Werby[/FONT]
    [URL="http://www.computersculpture.com/"]Website[/URL]


  5. #5
    Member
    Join Date
    Nov 2014
    Posts
    75
    Downloads
    0
    Uploads
    0

    Default

    Quote Originally Posted by awerby View Post
    Looking at the design of that machine: momus CNC | benchtop DIY router plans | home page it's not surprising that it was not very accurate. The idea of using a long timing belt instead of a leadscrew isn't that great for CNC cutting machines, although it seems to be a popular way of driving 3D printers. But printers don't have to contend with cutting forces. Replacing that system with leadscrews should help, although it's hard to say how much, since no specifications about the accuracy of those screws were included in that ad.
    I use one of these screws in the z axis i can say it is good quality and steel.the thing that i dont know for sure is if it is good diameter for cnc...ot has a lead of 8 mm per turn but it is 8mm thick.will it be ok for a cnc like this?or i need biggwe in diameter?i will also use diy antibackshlash nuts.Another question i would like to ask is if i can run a 4th stepper with the same dm controller..?or i will need to purchase another controller for the same axis?i want to put another moter on the other side of y axis(now it is driven by one)



  6. #6
    Member awerby's Avatar
    Join Date
    Apr 2004
    Posts
    5728
    Downloads
    0
    Uploads
    0

    Default Re: CNC improvements

    8mm seems a bit thin for a leadscrew. I'd suggest 12mm at minimum, and thicker would be better if they are longer than about 300mm. Are your anti-backlash nuts springloaded? Just be aware that if the spring force is exceeded, then you get all that inaccuracy back and more. To run dual motors, you'd need an extra driver, but not necessarily another controller (never heard of the dm, sorry.) Usually you'd use the extra motor on the longest axis, which by convention is called "X".

    [FONT=Verdana]Andrew Werby[/FONT]
    [URL="http://www.computersculpture.com/"]Website[/URL]


  7. #7
    Member
    Join Date
    Nov 2014
    Posts
    75
    Downloads
    0
    Uploads
    0

    Default Re: CNC improvements

    thanks awerby!The problem is that i cant find a 4 start lead screw with 12 mm diameter in Greece or on ebay....the ones i find are one start and have 2 mm lead the most i think...the nut is spring loaded delrin nut (two nuts jointed together and forced apart with a spring) i think when the spring force go away i will build another one ....i think this should hold for a year maybe?At this pont the machine is on +- 0.2 mm accuracy but i want to use it in building musical instruments and accuracy is needed...if you have any more helpful advice please tell me!



  8. #8
    Member awerby's Avatar
    Join Date
    Apr 2004
    Posts
    5728
    Downloads
    0
    Uploads
    0

    Default Re: CNC improvements

    I wasn't talking about the spring losing its force over time, but about it being overpowered by the resistance of the material you're cutting. If you have a spring with 1000 grams of power to resist compression but it runs into a cutting force of 1250 grams in the other direction, what do you think is going to happen?

    You don't need a 4-start screw, you just need a relatively fast one. But try to find one that specifies how accurate it is - cheap screws can be shockingly imprecise.

    [FONT=Verdana]Andrew Werby[/FONT]
    [URL="http://www.computersculpture.com/"]Website[/URL]


  9. #9
    Member
    Join Date
    Nov 2014
    Posts
    75
    Downloads
    0
    Uploads
    0

    Default

    Quote Originally Posted by awerby View Post
    I wasn't talking about the spring losing its force over time, but about it being overpowered by the resistance of the material you're cutting. If you have a spring with 1000 grams of power to resist compression but it runs into a cutting force of 1250 grams in the other direction, what do you think is going to happen?

    You don't need a 4-start screw, you just need a relatively fast one. But try to find one that specifies how accurate it is - cheap screws can be shockingly imprecise.
    Ohhh thanks alot!Now i see what you meant!I only have access on ebay to buy the screw...maybe a china ball screw is better?



  10. #10
    Member awerby's Avatar
    Join Date
    Apr 2004
    Posts
    5728
    Downloads
    0
    Uploads
    0

    Default Re: CNC improvements

    While there are good-quality ball screws made in China, there are also some made there which aren't suitable for use in precision machinery, and are only useful for opening skylights, etc. You need to look at the specs for the screw you're thinking of buying, particularly the expected accuracy expressed in +/- mm per 300mm of deviation from perfect (or in decimal inches per foot). The best grade of rolled ballscrews (C7) have an accuracy of +/-.003" per foot, which is okay for most wood projects, but not great for mechanical parts in metal. For better precision, you need to go with ground ball screws - or acme screws, which are typically more precise than ball screws. Here's an article (probably more than you want to know) about the differences between rolled and ground ball screws: Precision Ballscrew Threads: Rolled versus Ground

    [FONT=Verdana]Andrew Werby[/FONT]
    [URL="http://www.computersculpture.com/"]Website[/URL]


  11. #11
    Member
    Join Date
    Jan 2007
    Location
    usa
    Posts
    1795
    Downloads
    0
    Uploads
    0

    Default Re: CNC improvements

    for wood, plastic, and some light aluminum work the Chinese also works..
    for a metal milling machine realy apply what awerby telling

    for the present beltdrive, the ballscrew, any type would be better..

    you can buy on aliexpress if you haven't ebay account....

    CNC Linear SFU 1204 BallScrew Set 1pc RM1204 L 330mm + 1pc SFU1204 Ball Screw Anti Backlash BallNut with end machining-in Ball Screws from Industry & Business on Aliexpress.com | Alibaba Group

    for your machine I think you could get for each axis about 25 dollar a screw...


    they make you to size.. means any size you can buy.. under 350 mm the 12 dia screw works... over 350 mm distance use the 16 mm screw...



  12. #12
    Registered
    Join Date
    Jun 2014
    Posts
    777
    Downloads
    0
    Uploads
    0

    Default Re: CNC improvements

    SFU Ball screws are ok for hobby, the plastic ball return plugs wear if used extensively and you start finding broken ball bearings falling out. They do some sfu's in long lengths but really anything over a 1000mm you want more than a 5mm pitch imo.



  13. #13
    Member
    Join Date
    Nov 2014
    Posts
    75
    Downloads
    0
    Uploads
    0

    Default Re: CNC improvements

    So i am a little confused....i thought that the best accuracy is achieved with a ball screw...i can find lead screws in my town hat have a lead of 2 mm and i can make an antibackshlash nut from delrin pom but it will become really slow i guess....i want to thank you all for helping me !!



  14. #14
    Registered
    Join Date
    Jun 2014
    Posts
    777
    Downloads
    0
    Uploads
    0

    Default

    Quote Originally Posted by pikolo View Post
    So i am a little confused....i thought that the best accuracy is achieved with a ball screw...i can find lead screws in my town hat have a lead of 2 mm and i can make an antibackshlash nut from delrin pom but it will become really slow i guess....i want to thank you all for helping me !!
    Ball screws are more accurate than trapezodial screws which typically have a pitch error of 0.1mm. And yes it will be slower than ball screw



  15. #15
    Member
    Join Date
    Nov 2014
    Posts
    75
    Downloads
    0
    Uploads
    0

    Default Re: CNC improvements

    ah ok !So i will stick with the ball screws and see what happens....



  16. #16
    Member awerby's Avatar
    Join Date
    Apr 2004
    Posts
    5728
    Downloads
    0
    Uploads
    0

    Default Re: CNC improvements

    Quote Originally Posted by Jon.N.CNC View Post
    Ball screws are more accurate than trapezodial screws which typically have a pitch error of 0.1mm. And yes it will be slower than ball screw
    There's as much variation in ball screws as in any other kind. Some are shockingly inaccurate. There may be trapezoidal-thread screws with a pitch error of .1mm, but they also can be much more accurate than that. The "speed" of a screw depends on its number of threads per linear unit of measurement. You can find ball screws with fine pitches, and trapezoidal screws with coarse ones. Ball screws do present less resistance to motion than trapezoidal ones, but that's really the only generalization one can make about them.

    [FONT=Verdana]Andrew Werby[/FONT]
    [URL="http://www.computersculpture.com/"]Website[/URL]


  17. #17
    Member
    Join Date
    Apr 2009
    Location
    USA
    Posts
    5516
    Downloads
    0
    Uploads
    0

    Default Re: CNC improvements

    You can find a precision ground ACME screw or trapezoidal screw that is just as accurate as a ballscrew. The problem is they can almost cost as much if you buy new, and if you use a fine pitch you will be limited in speed.

    The effective lead of any screw is the pitch times the number of leads. Yes even ballscrews can have multiple leads. A typical rolled ballscrew has an accuracy grade of C7 which is about 50 microns or about .002" over 300mm/1ft. But they will also have a typical backlash of about the same amount because of the clearance needed to overcome any variation in pitch that would cause the scrw to otherwise bind. The better solution for rolled ballscrews would be a double nut that is preloaded. A C5 precision ground ballscrew is half that over the same length, and a C3 half that of a C5. Also because of the more precise process, the ballnuts of precision ground ballscrews can use other ways to achieve anti-backlash, like skewing the ball races to achieve a preload. They also make a double nut for higher load capacities. Note that anything below C3 is pretty rare to find, and will cost way more.



  18. #18
    Member
    Join Date
    Apr 2013
    Location
    Sweden
    Posts
    1899
    Downloads
    2
    Uploads
    0

    Default Re: CNC improvements

    Quote Originally Posted by pikolo View Post
    Hello i would like to asj somethibg else...the first problen was solved it was a faulty stretched belt.But i stikl need more accuracy tha 0.2 mm... i am thinking of changing the belts with this kind of lead screw 3D Printer 8mm Lead Screw Rod 300mm - 4 Start Z Axis 30cm Linear Rail Bar Shaft | eBay and an antibackshlash nut. Whats your opinion on this?thanks!if others can help please do!
    You will not be happy with that. Get a proper ball screw and you will never look back.



  19. #19
    Registered
    Join Date
    Mar 2013
    Location
    USA
    Posts
    209
    Downloads
    0
    Uploads
    0

    Default Re: CNC improvements

    Hi!
    I tend to disagree with the ball screw recommendation, because of the experience.
    There's nothing wrong with the original Momus design, using properly tensioned cevlar reinvorced belts on the x and y - the beauty is that there can't be any backlash.
    The 2nd X stepper motor is a must to make this work, which isn't in the original manual, unfortunately...

    With that mod, the long version that I've been running for a few years now has no problems holding 0.001" precision on the X and Y.
    The machine I've built back then should be pretty worn out by now, but it isn't - and I've done nothing special to it, except for keeping the rails clean - using WD-40.
    So far, several hundred board foot of poplar boards were profiled into smaller parts without any adjustments and it still holds up to cut engravings into jewelry without any issues..

    Long story short, stick to the belts - it really works.
    --
    Mac



  20. #20
    Member
    Join Date
    Apr 2013
    Location
    Sweden
    Posts
    1899
    Downloads
    2
    Uploads
    0

    Default Re: CNC improvements

    Quote Originally Posted by Mac.CNC View Post
    Hi!
    I tend to disagree with the ball screw recommendation, because of the experience.
    There's nothing wrong with the original Momus design, using properly tensioned cevlar reinvorced belts on the x and y - the beauty is that there can't be any backlash.
    The 2nd X stepper motor is a must to make this work, which isn't in the original manual, unfortunately...

    With that mod, the long version that I've been running for a few years now has no problems holding 0.001" precision on the X and Y.
    The machine I've built back then should be pretty worn out by now, but it isn't - and I've done nothing special to it, except for keeping the rails clean - using WD-40.
    So far, several hundred board foot of poplar boards were profiled into smaller parts without any adjustments and it still holds up to cut engravings into jewelry without any issues..

    Long story short, stick to the belts - it really works.
    --
    Mac
    0.001" = 0.0254mm

    Assuming it is per meter... that is good for wood or engraving, but with ball screws you can get better than that.

    Another thing... don't assume that there is no back lash, unless you have especially made your pulleys for your belts. I think cheap Chinese pulleys have always backlash and I think if you have a long belt you will always have tension caused backlash unless your belt is very very strong. The greatest advantage of belt drive is that it is quiet and can be very fast. Also that it needs no lubrication and I don't think it needs that much care at all. Ball screws must be protected, regularly greased and cleaned and makes more noise, but no doubt that they are more accurate, unless you are using really high quality things. Never the less, if the goal is 0.2mm or better accuracy than a belt drive may also be just as good. 0.2mm is a quite low set requirement.



Page 1 of 3 123 LastLast

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  


About CNCzone.com

    We are the largest and most active discussion forum for manufacturing industry. The site is 100% free to join and use, so join today!

Follow us on


Our Brands

CNC improvements

CNC improvements