Hi Jörg, the accuracy of the servo is one part of the (CNC) game. I found on my machine that the (chinese) ball screws have non negligible tolerances as well (OK, man kann es auch Asymptoten-Polieren nennen...). In my case this was up to 0,25 %. This sounds not much but 100 mm can deviate by 0,25 mm... I used a low tolerance meter scale (DIN866 Form A, 500 mm and 1000 mm) in combination with my reference camera to adjust the steps/mm on each axis. This works quite well and is reasonably priced (38 € for DIN866/1000 and 6 € for endoscope camera). And my (cheap chinese) end mills have tolerances as well. My 1/8" end mills have a "diameter range" from 3,07...3,17 mm and will bring futher 0,1 mm tolerance into the system (if not considered in the tool data base). Grüße aus Baden-Württemberg Kassen Oldewurtel
Hey Kassen, I Totally agree, the arrived accuracy is a combination of many factors - machine deflection is another big one to add. I think most diy machines have C5 class ball screws and yes the tolerance over 300mm is not good. Backlash from the fixed bearing and lost motion due to high pre-loads of the linear guide system are also at play. Anyways I always wanted to measure out the positioning accuracy as Eding CNC allows for a compensation. I looked at a glass scale DRO that can be purchased relatively inexpensive and temporarily attach it. The camera is a great idea ! Thank you for that suggestion I will check it out. Und Dir noch einen schönen Tag !
I am currently traveling and do not have the exact model number. However it is a “Digital Dial Indicator” this particular model has a large measuring range I think it is 50mm or so and a resolution of 0.01mm.
I think it is the same. The only difference I saw is that I was able to take the backlash compensation on x and y down. Less than half. I think that has to do with the accuracy of step positions in micro stepping of the stepper system. The part accuracy however is equal.
@@JBWorx had to go cheap... thus only 50mm thick , machine footprint is 1000mm x 1000mm x220mm xyz from stone suppliers running on 20mm linear rail comes from guys who makes granite counters ...sides is rollbolted and anchor glue on 3 sides then raised sides is diamond scraped to be flat.
Servos are not going to make a more accurate machine. Steppers are just as accurate. Your money is best spent in better ball screws/linear guides, and accurate dimensions for your machine, the steppers control linear actuation, they are just as accurate as servos, just not as fast or quiet.
Please do not stop making them, I am learning a lot and enjoying it very much.
That is a great motivation, thank you !
Hi Jörg,
the accuracy of the servo is one part of the (CNC) game.
I found on my machine that the (chinese) ball screws have non negligible tolerances as well (OK, man kann es auch Asymptoten-Polieren nennen...).
In my case this was up to 0,25 %. This sounds not much but 100 mm can deviate by 0,25 mm...
I used a low tolerance meter scale (DIN866 Form A, 500 mm and 1000 mm) in combination with my reference camera to adjust the steps/mm on each axis. This works quite well and is reasonably priced (38 € for DIN866/1000 and 6 € for endoscope camera).
And my (cheap chinese) end mills have tolerances as well. My 1/8" end mills have a "diameter range" from 3,07...3,17 mm and will bring futher 0,1 mm tolerance into the system (if not considered in the tool data base).
Grüße aus Baden-Württemberg
Kassen Oldewurtel
Hey Kassen,
I Totally agree, the arrived accuracy is a combination of many factors - machine deflection is another big one to add. I think most diy machines have C5 class ball screws and yes the tolerance over 300mm is not good. Backlash from the fixed bearing and lost motion due to high pre-loads of the linear guide system are also at play. Anyways I always wanted to measure out the positioning accuracy as Eding CNC allows for a compensation. I looked at a glass scale DRO that can be purchased relatively inexpensive and temporarily attach it. The camera is a great idea ! Thank you for that suggestion I will check it out. Und Dir noch einen schönen Tag !
Good job
Thank you !
Thank you!
Excellent thanks
👍
Great Video - Thank You! :)
Glad you liked it
hello, friend, whats the name of the mitutoyo meter you use?
I am currently traveling and do not have the exact model number. However it is a “Digital Dial Indicator” this particular model has a large measuring range I think it is 50mm or so and a resolution of 0.01mm.
is your system more accurate with servos in stead of steppers regardless of speed benefits ?
I think it is the same. The only difference I saw is that I was able to take the backlash compensation on x and y down. Less than half. I think that has to do with the accuracy of step positions in micro stepping of the stepper system. The part accuracy however is equal.
@@JBWorx Thank you , much appreciated. i have a converted bridgeport and busy building a granite based cnc router based of piotr fox `s machine.
@@JBWorx may i enquire as to how much microstepping you used and your accelerations ...on your previous stepper setup .thnx
Oh wow that is a nice machine !!! Where do you gut the granite from ?
@@JBWorx had to go cheap... thus only 50mm thick , machine footprint is 1000mm x 1000mm x220mm xyz from stone suppliers running on 20mm linear rail comes from guys who makes granite counters ...sides is rollbolted and anchor glue on 3 sides then raised sides is diamond scraped to be flat.
Servos are not going to make a more accurate machine. Steppers are just as accurate. Your money is best spent in better ball screws/linear guides, and accurate dimensions for your machine, the steppers control linear actuation, they are just as accurate as servos, just not as fast or quiet.
But untuned servos can be way less accurate hence the test after tuning.