Dynomotion

Group: DynoMotion Message: 14562 From: bordaco Date: 4/5/2017
Subject: Backlash problems...or not ??
Hello

I having a problems cutting circles, but first let me explain my setup:
Kmotion+Kanalog, DC servo drive in all axes, with they velocity control, position encoder in the Z axis screw, and linear glass scales (heidenhain) in X and Y.

The system works fine, smooth even at high speed.

But now I need to make some very precise holes in steel, and i was making some trials to check the the cuts,
By example a 40mm diameter hole, when i measure in the X and Y directions a got exact measurements, but when I check the diagonals, i got +0.1mm in one diagonal and -0.1 in the opposite diagonal.

I can pass a error of 0.03 mm maximum, but 0.1 its too much.


So i tried to correct with the backlash but, the error in X and Y is almost 0.01mm, in linear G1 movement. And any backlash compensation only wake worst.


Any suggestion?


Thanks
Group: DynoMotion Message: 14563 From: Tom Kerekes Date: 4/5/2017
Subject: Re: Backlash problems...or not ??

Could it be an orthogonality issue?

https://goo.gl/images/VgsCEw

Regards

TK


On 4/5/2017 2:55 PM, bordaco@... [DynoMotion] wrote:
 

Hello

I having a problems cutting circles, but first let me explain my setup:
Kmotion+Kanalog, DC servo drive in all axes, with they velocity control, position encoder in the Z axis screw, and linear glass scales (heidenhain) in X and Y.

The system works fine, smooth even at high speed.

But now I need to make some very precise holes in steel, and i was making some trials to check the the cuts,
By example a 40mm diameter hole, when i measure in the X and Y directions a got exact measurements, but when I check the diagonals, i got +0.1mm in one diagonal and -0.1 in the opposite diagonal.

I can pass a error of 0.03 mm maximum, but 0.1 its too much.


So i tried to correct with the backlash but, the error in X and Y is almost 0.01mm, in linear G1 movement. And any backlash compensation only wake worst.


Any suggestion?


Thanks
Group: DynoMotion Message: 14585 From: bordaco Date: 4/13/2017
Subject: Re: Backlash problems...or not ??
Dear Tom

I have been trying to solve this issue, physically the machine does not have (or at least I have not found) a way to correct the displacement of some of the axes, the bench is quite robust.

So try to solve it with geo correction table, and achieve an almost perfect circle, the deviation is 0.01mm that is fine.

But now I have other problems, when I home the exes (x and y) the position readout shows me a displacement of 6.604mm in both axes. (Not much problem but the operator no longer trust the machine)

When I cut a rectangle I now have a displacement of 0.1mm on one side, it's a trapezium.

Also, we were affected to the compensation of the diameter of the tool, for the tests we used an end mill of 20mm diameter, I usually declare it to 19.6mm because is a used tool, but after using the geo correction table I now have to declare 18.325mm to get The final cut measure

Also the work coordinate is shifted (G55 ... G59.)

When the geo correction table is disabled, the square cut is accurate, the tool set is the same as usual, and the working set is maintained, but the circle cut has problems.

Unfortunately my CAM program only reduces G2 and G3 cuts to quadrants, and the truth is I do not know if the system is converting the arcs into segments, because otherwise the control of the start and end points of each segment would not allow to much deviation of the resulting arc.

Any suggestion?

this is my GEo table

2,2
28.8,28.8
-14.4, -14.4
0,0,-14.19,-14.19
0,1,14.09,14.09
1,0,-14.09,14.09
1,1,14.19,14.19

Thanks for the help

Boris B.


---In DynoMotion@yahoogroups.com, <tk@...> wrote :

Could it be an orthogonality issue?

https://goo.gl/images/VgsCEw

Regards

TK


On 4/5/2017 2:55 PM, bordaco@... [DynoMotion] wrote:
 

Hello

I having a problems cutting circles, but first let me explain my setup:
Kmotion+Kanalog, DC servo drive in all axes, with they velocity control, position encoder in the Z axis screw, and linear glass scales (heidenhain) in X and Y.

The system works fine, smooth even at high speed.

But now I need to make some very precise holes in steel, and i was making some trials to check the the cuts,
By example a 40mm diameter hole, when i measure in the X and Y directions a got exact measurements, but when I check the diagonals, i got +0.1mm in one diagonal and -0.1 in the opposite diagonal.

I can pass a error of 0.03 mm maximum, but 0.1 its too much.


So i tried to correct with the backlash but, the error in X and Y is almost 0.01mm, in linear G1 movement. And any backlash compensation only wake worst.


Any suggestion?


Thanks
Group: DynoMotion Message: 14586 From: Tom Kerekes Date: 4/13/2017
Subject: Re: Backlash problems...or not ??
Hi Boris,

I think you have some typos in the table. Is that what you are actually using ?

The signs should be:

- -
+ -
- +
+ +

If you want to change only the orthogonality and not scales or trapezoidal then for example start with your 4 point square with 28.8 spacing. Then keeping all spacings the same add skew by shifting the two top points left or right by say 0.1

Note the table is in inches. Although the size shouldn't matter only the angle it is skewed. 

There is an option in the Tool Setup | Trajectory Planner to change Arcs to line Segments. 

Please try those changes and see where you stand. 

Regards
TK

On Apr 13, 2017, at 11:26 AM, bordaco@... [DynoMotion] <DynoMotion@yahoogroups.com> wrote:

 

Dear Tom

I have been trying to solve this issue, physically the machine does not have (or at least I have not found) a way to correct the displacement of some of the axes, the bench is quite robust.

So try to solve it with geo correction table, and achieve an almost perfect circle, the deviation is 0.01mm that is fine.

But now I have other problems, when I home the exes (x and y) the position readout shows me a displacement of 6.604mm in both axes. (Not much problem but the operator no longer trust the machine)

When I cut a rectangle I now have a displacement of 0.1mm on one side, it's a trapezium.

Also, we were affected to the compensation of the diameter of the tool, for the tests we used an end mill of 20mm diameter, I usually declare it to 19.6mm because is a used tool, but after using the geo correction table I now have to declare 18.325mm to get The final cut measure

Also the work coordinate is shifted (G55 ... G59.)

When the geo correction table is disabled, the square cut is accurate, the tool set is the same as usual, and the working set is maintained, but the circle cut has problems.

Unfortunately my CAM program only reduces G2 and G3 cuts to quadrants, and the truth is I do not know if the system is converting the arcs into segments, because otherwise the control of the start and end points of each segment would not allow to much deviation of the resulting arc.

Any suggestion?

this is my GEo table

2,2
28.8,28.8
-14.4, -14.4
0,0,-14.19,-14.19
0,1,14.09,14.09
1,0,-14.09,14.09
1,1,14.19,14.19

Thanks for the help

Boris B.


---In DynoMotion@yahoogroups.com, <tk@...> wrote :

Could it be an orthogonality issue?

https://goo.gl/images/VgsCEw

Regards

TK


On 4/5/2017 2:55 PM, bordaco@... [DynoMotion] wrote:
 

Hello

I having a problems cutting circles, but first let me explain my setup:
Kmotion+Kanalog, DC servo drive in all axes, with they velocity control, position encoder in the Z axis screw, and linear glass scales (heidenhain) in X and Y.

The system works fine, smooth even at high speed.

But now I need to make some very precise holes in steel, and i was making some trials to check the the cuts,
By example a 40mm diameter hole, when i measure in the X and Y directions a got exact measurements, but when I check the diagonals, i got +0.1mm in one diagonal and -0.1 in the opposite diagonal.

I can pass a error of 0.03 mm maximum, but 0.1 its too much.


So i tried to correct with the backlash but, the error in X and Y is almost 0.01mm, in linear G1 movement. And any backlash compensation only wake worst.


Any suggestion?


Thanks