Dynomotion

Group: DynoMotion Message: 13351 From: Colin Fera Date: 6/12/2016
Subject: backlash compensation
Hi Tom,

I measured the backlash on the CNC mill that runs of the kflop and am seeing about .00075" to .001" of backlash in all 3 axis's.  Actually I am not sure if backlash is the right word because I cannot push agains the axis externally and take anywhere near this much up. 

So I want to put in backlash compensation and this amount equates to about 30 counts.  Thats simple enough.  

Do you have any recommendations on the speed?

I have not tried it with coordinated movement but with step jogging I have noticed that the backlash compensation is audible when the speed is set high.

Thanks,
Colin
Group: DynoMotion Message: 13369 From: Colin Fera Date: 6/16/2016
Subject: Re: backlash compensation
Anyone have any ideas on the speed setting for the backlash compensation?

Thanks,
Colin

On Sun, Jun 12, 2016 at 5:41 PM, Colin Fera <colin.fera@...> wrote:
Hi Tom,

I measured the backlash on the CNC mill that runs of the kflop and am seeing about .00075" to .001" of backlash in all 3 axis's.  Actually I am not sure if backlash is the right word because I cannot push agains the axis externally and take anywhere near this much up. 

So I want to put in backlash compensation and this amount equates to about 30 counts.  Thats simple enough.  

Do you have any recommendations on the speed?

I have not tried it with coordinated movement but with step jogging I have noticed that the backlash compensation is audible when the speed is set high.

Thanks,
Colin
Group: DynoMotion Message: 13371 From: Tom Kerekes Date: 6/16/2016
Subject: Re: backlash compensation

Hi Colin,

I think you will need to find the rate that works best for your machine empirically.  Too slow and the backlash wont be corrected fast enough and leave a visible error.  Too fast and possible jerk and over shoot might occur causing a visible error.  I'd start with correcting the error in ~ 50 milliseconds.  So the rate to correct 30 counts in 50 milliseconds would be:

30 counts / 0.05sec = 600 counts/sec

I don't know why it would be "audible" at high speed.  What speed were you using?  The effect should only occur when the direction reverses.

What Version are you running?

What type of system do you have?

Regards

TK



On 6/16/2016 6:31 PM, Colin Fera colin.fera@... [DynoMotion] wrote:
 
Anyone have any ideas on the speed setting for the backlash compensation?

Thanks,
Colin

On Sun, Jun 12, 2016 at 5:41 PM, Colin Fera <colin.fera@...> wrote:
Hi Tom,

I measured the backlash on the CNC mill that runs of the kflop and am seeing about .00075" to .001" of backlash in all 3 axis's.  Actually I am not sure if backlash is the right word because I cannot push agains the axis externally and take anywhere near this much up. 

So I want to put in backlash compensation and this amount equates to about 30 counts.  Thats simple enough.  

Do you have any recommendations on the speed?

I have not tried it with coordinated movement but with step jogging I have noticed that the backlash compensation is audible when the speed is set high.

Thanks,
Colin
Group: DynoMotion Message: 13372 From: Colin Fera Date: 6/16/2016
Subject: Re: backlash compensation
Hi Tom,

I tried 100,000 (and got a loud click when reversing).  10,000 (some noise), 5000 no-noise, and 2000.    

I am running 4.33.

I assumed the noise was a sudden acceleration tacking up the backlash.  
 
The system is a 3 axis mill with 1" ground ball screws that are belt driven from 750 watt servos.  The machine itself shows no visible wear and still has chrome plated ways and the ball screws look good visibly.   I think the claimed accuracy when it was new was  .001".


Thanks,
Colin




On Thu, Jun 16, 2016 at 6:48 PM, Tom Kerekes tk@... [DynoMotion] <DynoMotion@yahoogroups.com> wrote:
 

Hi Colin,

I think you will need to find the rate that works best for your machine empirically.  Too slow and the backlash wont be corrected fast enough and leave a visible error.  Too fast and possible jerk and over shoot might occur causing a visible error.  I'd start with correcting the error in ~ 50 milliseconds.  So the rate to correct 30 counts in 50 milliseconds would be:

30 counts / 0.05sec = 600 counts/sec

I don't know why it would be "audible" at high speed.  What speed were you using?  The effect should only occur when the direction reverses.

What Version are you running?

What type of system do you have?

Regards

TK



On 6/16/2016 6:31 PM, Colin Fera colin.fera@... [DynoMotion] wrote:
 
Anyone have any ideas on the speed setting for the backlash compensation?

Thanks,
Colin

On Sun, Jun 12, 2016 at 5:41 PM, Colin Fera <colin.fera@...> wrote:
Hi Tom,

I measured the backlash on the CNC mill that runs of the kflop and am seeing about .00075" to .001" of backlash in all 3 axis's.  Actually I am not sure if backlash is the right word because I cannot push agains the axis externally and take anywhere near this much up. 

So I want to put in backlash compensation and this amount equates to about 30 counts.  Thats simple enough.  

Do you have any recommendations on the speed?

I have not tried it with coordinated movement but with step jogging I have noticed that the backlash compensation is audible when the speed is set high.

Thanks,
Colin