Jittery Random Operation, CNC Router

[hkinkade]

Well unfortunately I don't have many details, but here goes.

*The machine is an old K2 model, but the drive units have all been redone with leadshine servo and servo drives.

*Three axis with ball screw on Z, racks on X and Y, and dual drives on each leg of of the Y axis.

*Mach 3 controller

*CAM via fusion 360

I was 'cutting air' prior to cutting a profile for a planer stand I'd designed. I heard an occasional rough moment during this operation. I repeated several times and rough movement occured randomly in different areas of the operation. I considered VFD noise/interference then turned the spindle vfd completly off (at the breaker). Still the problem would reoccure. My next step is to redo the wiring, as the company who had the unit before me did a poor job of wiring the retrofit. I have run the unit a fair amount over the last few months without any problems.

Here's my ask, what else should I consider checking, grounding . . .????

Howard

 

[memphisjed]

*Mach 3 controller

first place to look.

 

[hkinkade]

I did give some consideration to my Mach 3 setup ( constant velocity in particular). But, the problem occurs randomly and in a different spot every time. Once in a simple retract of the Z axis. Did you have something specifically I should consider with Mach 3?

 

[CarbideBob]

*The machine is an old K2 model, but the drive units have all been redone with leadshine servo and servo drives.

Hmm. seems not a Mach 3 problem.

 

[memphisjed]

Hmm. seems not a Mach 3 problem.

If mach straight to leadshines or mach to smoothstepper(et al) are 2 different things.
Random and jittery makes my first guess communication stutter step.

 

[hkinkade]

Additional information - Mach 3 interfaces with POKEYS57 breakout board.

Jed, please expand on what 'communicaiton stutter step' means.

​

 

[hkinkade]

Please view attached link. You can see and hear the stutter / jitter as the program starts. The Z axis is raising from G54 position to begin the process.

CNC Jitter Video

 

[memphisjed]

Please view attached link. You can see and hear the stutter / jitter as the program starts. The Z axis is raising from G54 position to begin the process.

CNC Jitter Video

sounds mechanical? could still be communication. Are you running a smoothstepper or another intermediate card?
If you move z up and down does it make the same ugh motion in relatively the same spots?
Slow speed, medium speed, hyperspeed?

 

[gustafson]

Is it always in the Z?

 

[hkinkade]

Mach3 communicates with CNC controller, PoKeys5, which controls the servo drives. It has also occured during X/Y motion without Z motion. The location of the problem is random in location/time. I have done a brief mechanical check and found nothing apparent. I'll do a more thorough check today. I'll also do manual traverses to see if I can get it to repeat. While I'm at it I'll run some older programs and see if it's unique to this file or not.

 

[CarbideBob]

At it's base in the world Mach is a step and direction guy.
Put on analog world servos. Now you ask the drive to do this conversion. This adds a layer and possible problems.
Being random I would suspect too high in the "I" term at the amp but that wild ass guess.
Moving in a notch filter and making the "D" guy calm down may do the trick if you amp understands such but this is complicated to say the least.
Welcome to what should work but does some things bad that seem weird.

 

[hkinkade]

I want to thank everyone for the good input and suggestions. I figured it out, well kind of. I noticed the PC was sluggish getting starting and hung when I logged in. That seemed strange, so I shut it down and did a reboot. Problem solved! Not certain why, but it's always easy to blame microsoft or windows when you don't know what happened ;-)

 

[usolutions]

I have similar on my VMC, that uses RS232 drip feed. Do you use smoothing? otherwise it loads gcode with 1 micron movements that (in my case can not be drip fed fast enough) in your case controller might struggle to process trajectory on these 1000 lines to move 1 mm gcode sections

 

[hkinkade]

I have similar on my VMC, that uses RS232 drip feed. Do you use smoothing? otherwise it loads gcode with 1 micron movements that (in my case can not be drip fed fast enough) in your case controller might struggle to process trajectory on these 1000 lines to move 1 mm gcode sections

Problem was resolved with computer reboot. Not certain of root cause though.

I’m just beginning my learning curve on the design and operation of my machine. I’m a ‘hobbyist’ and recently purchased the machine. I have had to rework the poorly done retrofit to servo motors. The focus has been on the mechanicals. Next I’m going to go through the wiring and grounding of the system.

I’m not certain what drip feed is. Maybe you can enlighten me ( one more step on my learning curve climb ;-)

 

[usolutions]

Problem was resolved with computer reboot. Not certain of root cause though.

I’m just beginning my learning curve on the design and operation of my machine. I’m a ‘hobbyist’ and recently purchased the machine. I have had to rework the poorly done retrofit to servo motors. The focus has been on the mechanicals. Next I’m going to go through the wiring and grounding of the system.

I’m not certain what drip feed is. Maybe you can enlighten me ( one more step on my learning curve climb ;-)

I am not a pro myself, but I got 1997 Matsuura VMC, that has like 56kb of internal memory, so I have to feed gcode line by line from host PC to CNC control. If I leave adaptive as is by default, while doing curvature there are many many many line with 0.001mm difference in coordinates. However adaptive is roughing procedure, you can smooth out trajectory resulting in more broad movement at a cost of lesser precision, but it does not matter due to it being a roughing toolpath. Having supersmall movements with slow processing is movement already done, but next move is still in processing.