Please help. Y-axis drifting

[Andros]

I have an X2, cncfusion kit ballscrew kit, mach3, g540 controller, 48Volt Nema23 motors. Backlash adjusted and properly setup Steps per. The machine is 2.4ghz, with has been tested with the driver check and passed.

Created this is meshcam with roughing, xy finish, AND a pencil clean up.

But something happens in the Y axis. Check the pictures and see how the pencil pass cleans up the front of the oval but has drifted past the rear oval shape and leaves the rough edges, literally pencil passing in the air far past. I looks like it has come over a millimeter off?

The motors sound fine i think.

Why is this happening? Am i missing something here? Too tight gib screws? Missing steps?

This is what it sounds and looks like:
http://www.youtube.com/watch?v=lx5FjR5Lk7I

Front side is finished with pencil cleanup:


Back side is not because the pencil path has drifted too far from the surface:





Sent from my iPhone using Tapatalk - now Free

 

[Richard King]

If I were you I would remove part and put a piece of glass or plex-i-glass, stable piece of aluminum on the table and remove the cutter and put a fine tip pen or pencil in the spindle. Program your control to draw a circle the same size of your cut. on graph paper that you have taped to the glass. This will show you if it is the control or a mechanical problem. You should also be sure to turn off the spindle. I would also check the back lash in your balls screws or feed screws. Check the gibs as shown in the pictures. I would do a push pull check on both end of the saddle and the table. Be sure to mount the mag base on the base to check the saddle and put the mag base on the saddle to check the table. If not you are checking both errors and not one. I would Push the slide and let go and see if it springs back and that is your zero, then pull the same part and let go. see if there is any spring. check both ends of the gib and they should have the same slop on both end. Many times I loosen the gib first so it is not jammed to deep and you get no movement. On a small machine that size I would set the clearance at .001" or less, you need room for way lube. You may also want to check the ball screw nut mount to see if it is tight. The end brackets. etc. I would also test the Z axis too for clearance. Watch the drive belts and see if there is a lot of back lash in Y axis when you change directions. Be a detective. Good luck and keep us posted. Richard

PS the test is in the pictures are being done of a VMC column, but the technique is the same on all applications. I would use a .0005" indicator either a plunger or a stem type.

 

[DMS]

Based on what you describe, it sounds like you are missing steps on the y axis. If it's only occasional, it is probably happening during a period of heavy cutting, or of high acceleration. You can try slowing the feed rate down and see if it helps. If it happens again, stop the machine and move the spindle back to 0,0 and see if it really goes to 0,0. If it doesn't, then you have missed steps somewhere.

 

[Andros]

Thank you both.

Dms: move the spindle back to zero. Like using the go to zero command? And then visually check if the start point of the cut is really where the spindle goes to or?


Sent from my iPhone using Tapatalk - now Free

 

[DMS]

Exactly, just "go to zero" (the g code would be "g0 x0 y0", but make sure your tool is clear first). It would to see the error if your zero is on the corner of the part. 1mm is a fair amount, but if you are trying to determine if something has moved based on an arbitrary point in space, you are going to have trouble.

 

[bloomingtonmike]

Also realize you may need load (bit in material) to uncover a lost step error. My guess is you have a coupler slipping under load.

- - - Updated - - -

Based on what you describe, it sounds like you are missing steps on the y axis. If it's only occasional, it is probably happening during a period of heavy cutting, or of high acceleration. You can try slowing the feed rate down and see if it helps. If it happens again, stop the machine and move the spindle back to 0,0 and see if it really goes to 0,0. If it doesn't, then you have missed steps somewhere.

+1

Mark your coupler and shafts and see if they slip.

 

[Andros]

Hey dudes. I really appreciate all your input! Will try it out on wednesday evening. Ill keep you posted!


Sent from my iPhone using Tapatalk - now Free

 

[DAN_IN_MN]

I have an X2, cncfusion kit ballscrew kit, mach3, g540 controller, 48Volt Nema23 motors. Backlash adjusted and properly setup Steps per. The machine is 2.4ghz, with has been tested with the driver check and passed.

Created this is meshcam with roughing, xy finish, AND a pencil clean up.

But something happens in the Y axis. Check the pictures and see how the pencil pass cleans up the front of the oval but has drifted past the rear oval shape and leaves the rough edges, literally pencil passing in the air far past. I looks like it has come over a millimeter off?

The motors sound fine i think.

Why is this happening? Am i missing something here? Too tight gib screws? Missing steps?

This is what it sounds and looks like:
http://www.youtube.com/watch?v=lx5FjR5Lk7I

Front side is finished with pencil cleanup:


Back side is not because the pencil path has drifted too far from the surface:



.......

I take it that this is a new conversion and you're working out the bugs? I searched 48Volt Nema23 motor and found that they are stepper motors. If my memory serves me correct, stepper motors can loose steps if there is any binding. I've learned this working with a stepper motor equipped plasma table. The faster a motor goes, the less torque it has. The stepper motors I've used are a no feedback system. The machine doesn't know if something is jammed.

Like others stated, check couplings, but also check ball screw alignment and for their smooth operation.

You could also slow things down to see if it still looses steps. Slowing a motor increases it's torque.

Are you running faster on your pencil cleanup?

Good luck!

 

[Andros]

Brand new machine im trying to get precise. Yes.


Sent from my iPhone using Tapatalk - now Free

 

[Hawkeye]

When I first got my X2 running, it wouldn't run very consistently until Steve (jumps4) got me to slow the steppers down. Once I was starting to cut parts that mattered, I found that the bit ended up about 3/8" short of the start point in the X-axis. I went into Mach3 and slowed the motors down a bit more, with the accompanying velocity adjustment. That brought the accuracy back in line.

I haven't done an in-depth check of accuracy, but it's running to acceptable standards for now.