[davo727]

Hi, I am planning my mill conversion and wondered about pros and cons of the encoders reading directly on the end of the ballscrew as opposed to the servo. I noticed that Industrialhobbies has theres on the end of the ballscrews. One member expressed to me that he thought it was a bad idea and would cause the servodrive to pulse the servo back and forth excessively and burn it out. Anyone have any opinions or direct experience either way on this? Thanks, Dave

[Bubba]

On my mill/drill, I have BOTH types of setup. I originally put the encoder on the end of the ball screw as I was unsure of how much reduction would be necessary to make it work effectively. By putting the encoder on the screw, I would not have to make any configuration changes if I changed pulleys. Another factor was the servos did not have extensions on which to mount the encoders. The X axis still has the screw encoder. The Y axis was moved to the back end of the servo and in my configuration, this protects it from swarf and especially coolant as it is under the mill.
With a properly setup belt drive, there is no backlash in the belt so it will not hunt anymore than if it is on the servo. Also remember, you will be able to use a higher count encoder if it is on the screw as there is no reduction factor to take into account.
Point is, either will work. It depends on how you design the system.

[springlakecnc]

Hi Dave, encoders on ballscrew, no problem IF there is no backlash between the motor and encoder. A timing belt between the motor and ballscrew is ok, timing belts are no backlash, but if you have any kind of backlach in there, you will probably have an occolation problem with the servo motor, and never get it tuned. Only the high end cnc controllers can handle backlash between the encoder/scale and the servo motor, like on a huge boring mill where the cnc reads a scale for accuracy.

[davo727]

OK thanks for the info guys, Dave

[Paraprop]

Interesting issue you raised here.

I was just brainstorming about the limitations of my servo's when using the Gecko's 320 that I just received as normally the enoder used should be 250 or 500 cpr max to avoid issues at high feed speeds and also due to controller max step limitation.

My encoders are 1000 CPR. This means apparently that the geckos will see them with a resolution of 1000 x 4 = 4000 steps per rotation right ?

the servo's are 3000 RPM types.
the ball screw is a 5 TPI model.
I am using the DeskCNC controller which is limited to 125,000 SPS.

I guess I should be happy with a max feed of about 170 to 200 IPM based from what I could read in this forum. (not sure if the mass of my X-Y table and the size of my servo's will manag though)

Given that my servo's do not have much power (200 W) I will use a 3:1 belt reduction.

I will get 1000 cpr x 4 x 5 TPI = 20,000 SPI (steps per inch) with the encoder directly mounted onto the shaft of the ball screw.

My max ball screw speed will be 3000 / 3 = 1000 RPM

Therefore I can get 1000 RPM x 4 x 1000 cpr = 4,000,000 SPM (steps/Minute)

max feed will be: 4,000,000 / 20,000 = 200 IPM.

assuming a safety margin of 10% I'll get 180 IPM.

My max veloity will therefore be:

180 x 20,000 / 60 = 60,000 SPS (steps / second)

This means that with more powerfull servos, I could even increase the feed speed 2 fold with this configuration and still be within the limitations of the controller by using a redution of 1.5:1 or go in between with 2:1 like many do.

If I would use the same setup with the encoders on the servo shaft, the controller would have to provide 3 times more steps e.g. 180,000 with a reduction of 3:1 !!

Thanks for that great solution to my problem. you saved my the money for 3 new encoders !

Daniel.